{ "schema": "humaneness-bucket-agent-routes-v1", "updated": 1793050332.0078356, "complete_targets": 72, "total_targets": 72, "routes": [ { "target": "emo_Affection", "label": "Intense Affection", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Affection", "adapter_sha256": "c2525bde21a35ac6072ab6a39e762d9e974608a99767de4e5ec8778a7c44f8fc", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Affection.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Affection", "category": "emotion", "head": "emo_Affection", "pole": "high", "label": "Intense Affection", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-bbf5dbc0f2ea", "general": "Infuse the speech with the delicate texture of a shared secret, making the warmth feel intensely personal and slightly conspiratorial.", "inline": [ "Begin with a very soft breath support, almost whispering the initial sentiment.", "Maintain a close, intimate vocal proximity throughout, as if leaning in close to the listener.", "Let the final sound gently dissolve, leaving a lingering, sweet resonance in the air." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.4576562639080907, "target_raw": 0.5263361638784408, "ce": 6.249615201950073, "genuineness": 4.315049257278442, "identity": 0.5826258924603462, "wer": 0.004596273291925465, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.3609179707111778, "target_raw": 1.0174727001786232, "ce": 5.874443054199219, "genuineness": 4.497556467056274, "identity": 0.5263409471511841, "wer": 0.004718614718614718, "valid_fraction": 1, "identity_delta": -0.05628494530916214, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.2967382888120962, "target_raw": 0.9825758519768715, "ce": 6.0227724552154545, "genuineness": 4.463916902542114, "identity": 0.5440931805968284, "wer": 0.01038495448106432, "valid_fraction": 1, "identity_delta": -0.03853271186351776, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.1561816439949715, "target_raw": 0.9061500462889671, "ce": 6.140800580978394, "genuineness": 4.534961681365967, "identity": 0.6130444830656052, "wer": 0.018582721626199887, "valid_fraction": 0.98, "identity_delta": 0.030418590605258942, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.14055664481712465, "ci95_z": [ -0.04078124225805788, 0.2977539176145544 ], "language_deltas": { "en": -0.14921873526975615, "de": 0.2722656200328248, "fr": -0.028515644697419084, "nl": 0.33598634480357553, "es": 0.2722656392163981 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.11802812576293946, "delta_genuineness": -0.07104477882385254, "delta_wer": -0.008197767145135567, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.06417968189908169, "ci95_z": [ -0.3412695113384785, 0.16437499040821332 ], "language_deltas": { "en": -0.5624999725948951, "de": 0.08808596499074582, "fr": -0.040527346019485204, "nl": -0.1069336038984529, "es": 0.300976548026679 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.14832940101623535, "delta_genuineness": -0.033639564514160156, "delta_wer": 0.005666339762449602, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Amusement", "label": "Intense Amusement", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Amusement", "adapter_sha256": "b57747a62853673ca6ab449728a6e41bed834df8bc3b14a06e397a01dfc4469a", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Amusement.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Amusement", "category": "emotion", "head": "emo_Amusement", "pole": "high", "label": "Intense Amusement", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-eef73ce58f32", "general": "Focus on the physical release of amusement, allowing the voice to feel momentarily buoyant.", "inline": [ "Start with a slight vocal expansion, as if taking a delighted, slightly uneven breath.", "Allow the pace to subtly quicken, mirroring the internal rush of genuine laughter.", "Let the final sound trail off with a lingering, cheerful vocal brightness." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.24352050843170517, "target_raw": 0.6890353599190712, "ce": 6.229281940460205, "genuineness": 4.392244567871094, "identity": 0.5804609605669975, "wer": 0.00658008658008658, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.3007788049083724, "target_raw": 1.60637122631073, "ce": 6.281227035522461, "genuineness": 4.314923686981201, "identity": 0.529170643389225, "wer": 0.01662714097496706, "valid_fraction": 0.98, "identity_delta": -0.05129031717777252, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.594843747016008, "target_raw": 1.861518260240555, "ce": 6.241195955276489, "genuineness": 4.2461383295059205, "identity": 0.5557417982816696, "wer": 0.013120647468473554, "valid_fraction": 1, "identity_delta": -0.02471916228532791, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.6413671687856473, "target_raw": 1.9018845582008361, "ce": 6.252792282104492, "genuineness": 4.249150795936584, "identity": 0.5354432153701782, "wer": 0.011110979028599165, "valid_fraction": 1, "identity_delta": -0.04501774519681931, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.04652342176963932, "ci95_z": [ -0.08070311404080635, -0.023671855475246636 ], "language_deltas": { "en": -0.11113279240098912, "de": -0.03808593924423995, "fr": -0.03203125375682443, "nl": -0.036914016103218004, "es": -0.014453107342925075 }, "languages_improved": 0, "clear_gain": false, "delta_ce": -0.011596326828002929, "delta_genuineness": -0.0030124664306640627, "delta_wer": 0.002009668439874389, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.2940649421076357, "ci95_z": [ 0.1692968818374205, 0.41883300237785087 ], "language_deltas": { "en": 0.16347656998681448, "de": 0.48087159976044247, "fr": 0.27734373980290505, "nl": 0.12109376470528428, "es": 0.42753903628273215 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.04003108024597168, "delta_genuineness": -0.06878535747528076, "delta_wer": -0.0035064935064935067, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Anger", "label": "Intense Anger", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Anger", "adapter_sha256": "9bc3f92b6406d5965e79b08fe9a5733e070d4b0cec2720c3a1e2fb60683ebca8", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Anger.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Anger", "category": "emotion", "head": "emo_Anger", "pole": "high", "label": "Intense Anger", "floor": 3.0 }, "configuration": { "adapter": true, "scale": 0.75, "cfg": 2.5, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "manual-a407e8c9b493", "general": "Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Establish it from the first phrase. Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Sustain it through the thought. Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.", "Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.", "Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.46607421894866236, "target_raw": 0.09022372364997863, "ce": 6.247708940505982, "genuineness": 4.334253921508789, "identity": 0.5676545485854149, "wer": 0.010223978919631093, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.5969140636759205, "target_raw": 0.7069014951586723, "ce": 6.14636043548584, "genuineness": 4.002063121795654, "identity": 0.5741742491722107, "wer": 0.00935441370223979, "valid_fraction": 1, "identity_delta": 0.006519700586795807, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.9923828265026841, "target_raw": 0.9363271617889404, "ce": 6.125391759872437, "genuineness": 3.936980586051941, "identity": 0.5668306910991668, "wer": 0.018413989519252678, "valid_fraction": 0.98, "identity_delta": -0.0008238574862480163, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.5969140636759205, "target_raw": 0.7069014951586723, "ce": 6.14636043548584, "genuineness": 4.002063121795654, "identity": 0.5741742491722107, "wer": 0.00935441370223979, "valid_fraction": 1, "identity_delta": 0.006519700586795807, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.39546876282676363, "ci95_z": [ 0.14320313216789743, 0.6621875334876457 ], "language_deltas": { "en": 0.005859365809358935, "de": 0.17753907375753203, "fr": 0.8851562975986911, "nl": 0.4308593760836127, "es": 0.4779297008846235 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.02096867561340332, "delta_genuineness": -0.06508253574371338, "delta_wer": 0.009059575817012888, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.39546876282676363, "ci95_z": [ 0.14320313216789743, 0.6621875334876457 ], "language_deltas": { "en": 0.005859365809358935, "de": 0.17753907375753203, "fr": 0.8851562975986911, "nl": 0.4308593760836127, "es": 0.4779297008846235 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.02096867561340332, "delta_genuineness": -0.06508253574371338, "delta_wer": 0.009059575817012888, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Establish it from the first phrase. Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Sustain it through the thought. Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Unmistakable anger; hard consonants, forceful breath and outraged conviction. Keep every word clear. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "language": "de", "seed": 1702, "scale": 0.75, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "emo_Astonishment_Surprise", "label": "Intense Astonishment Surprise", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Astonishment_Surprise", "adapter_sha256": "f236a51ca66337597792ce50794cd4774bbdd888bf3a2853e85bd1a183953d41", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Astonishment_Surprise.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Astonishment_Surprise", "category": "emotion", "head": "emo_Astonishment_Surprise", "pole": "high", "label": "Intense Astonishment Surprise", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 3.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-bedaa97ab879", "general": "Target resonance shift; the surprise should sound as if the vocal tract has momentarily widened or constricted drastically, changing the perceived timbre.", "inline": [ "Deliver the opening with a tightly focused, forward placement in the mouth, low in resonance.", "Immediately shift the resonance upward and backward, creating a sudden, bright, almost metallic quality of surprise.", "Let the final articulation carry that heightened, bright resonance, sustaining the unusual timbre until the very end." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.2251367117254231, "target_raw": 0.4020830598473549, "ce": 6.257081289291381, "genuineness": 4.31404510974884, "identity": 0.5546397522091866, "wer": 0.008457745153397327, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 2.7873827526772526, "target_raw": 1.8169506388902663, "ce": 6.138417778015136, "genuineness": 4.11766776561737, "identity": 0.5250219690799713, "wer": 0.007492941840767927, "valid_fraction": 1, "identity_delta": -0.02961778312921524, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.804023387290719, "target_raw": 1.826139566898346, "ce": 6.15112603187561, "genuineness": 4.169947180747986, "identity": 0.5156369972229004, "wer": 0.006457745153397327, "valid_fraction": 1, "identity_delta": -0.039002754986286164, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 2.7873827526772526, "target_raw": 1.8169506388902663, "ce": 6.138417778015136, "genuineness": 4.11766776561737, "identity": 0.5250219690799713, "wer": 0.007492941840767927, "valid_fraction": 1, "identity_delta": -0.02961778312921524, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.016640634613466584, "ci95_z": [ -0.33966795072896655, 0.2641015536117598 ], "language_deltas": { "en": 0.004003902919018132, "de": 0.1722656467567928, "fr": 0.27490240543641065, "nl": -0.6671874372594812, "es": 0.2992186552145925 }, "languages_improved": 4, "clear_gain": false, "delta_ce": 0.012708253860473633, "delta_genuineness": 0.05227941513061524, "delta_wer": -0.0010351966873706003, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.016640634613466584, "ci95_z": [ -0.33966795072896655, 0.2641015536117598 ], "language_deltas": { "en": 0.004003902919018132, "de": 0.1722656467567928, "fr": 0.27490240543641065, "nl": -0.6671874372594812, "es": 0.2992186552145925 }, "languages_improved": 4, "clear_gain": false, "delta_ce": 0.012708253860473633, "delta_genuineness": 0.05227941513061524, "delta_wer": -0.0010351966873706003, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Target resonance shift; the surprise should sound as if the vocal tract has momentarily widened or constricted drastically, changing the perceived timbre.\nSCRIPT:\n(Deliver the opening with a tightly focused, forward placement in the mouth, low in resonance.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Immediately shift the resonance upward and backward, creating a sudden, bright, almost metallic quality of surprise.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Let the final articulation carry that heightened, bright resonance, sustaining the unusual timbre until the very end.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 3.0, "frames": 128, "reference": true } }, { "target": "emo_Awe", "label": "Intense Awe", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Awe", "adapter_sha256": "c6880c4233e759c8587f88c97d936beb798bbd36dc9cc2f8b3917fb889398024", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Awe.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Awe", "category": "emotion", "head": "emo_Awe", "pole": "high", "label": "Intense Awe", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-51b69a70d1d3", "general": "Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Establish it from the first phrase. Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Sustain it through the thought. Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Keep it audible in the final sentence. Let the local direction carry the intended delivery beside each sentence.", "inline": [ "Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Establish it from the first phrase. Let the local direction carry the intended delivery beside each sentence.", "Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Sustain it through the thought. Let the local direction carry the intended delivery beside each sentence.", "Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Keep it audible in the final sentence. Let the local direction carry the intended delivery beside each sentence." ], "cfg": 3, "recipe": "P09", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.2113085948620942, "target_raw": 0.2068780005723238, "ce": 6.273919725418091, "genuineness": 4.3545882654190065, "identity": 0.5631067588925361, "wer": 0.006967418546365915, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.6400684119082776, "target_raw": 0.8066306413710117, "ce": 6.142389135360718, "genuineness": 4.478226251602173, "identity": 0.5230773529410362, "wer": 0.010574534161490682, "valid_fraction": 1, "identity_delta": -0.04002940595149994, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.914980507545725, "target_raw": 0.9220309020578861, "ce": 6.0666160392761235, "genuineness": 4.436804404258728, "identity": 0.5165845075249672, "wer": 0.010848861283643893, "valid_fraction": 1, "identity_delta": -0.04652225136756897, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.8679101807188456, "target_raw": 0.9022721207141876, "ce": 6.058117361068725, "genuineness": 4.417474193572998, "identity": 0.5263149186968803, "wer": 0.011887351778656127, "valid_fraction": 1, "identity_delta": -0.036791840195655824, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.04707032682687929, "ci95_z": [ -0.22337888636819633, 0.30080079692193085 ], "language_deltas": { "en": 0.1895507722506342, "de": 0.30390626320091857, "fr": 0.3533203429785915, "nl": -0.20673826892267155, "es": -0.40468747537307626 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.008498678207397461, "delta_genuineness": 0.01933021068572998, "delta_wer": -0.001038490495012234, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.2749120956374473, "ci95_z": [ 0.1918456827786681, 0.35797850849622653 ], "language_deltas": { "en": 0.37499999645017307, "de": 0.3976562481141544, "fr": 0.1948242104439597, "nl": 0.24458005335247773, "es": 0.16249996982647164 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.07577309608459473, "delta_genuineness": -0.041421847343444826, "delta_wer": 0.00027432712215320914, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Establish it from the first phrase. Let the local direction carry the intended delivery beside each sentence.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Sustain it through the thought. Let the local direction carry the intended delivery beside each sentence.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Quiet astonishment at something vast and beautiful; open resonance, reverent breath and spacious emphasis. Keep it audible in the final sentence. Let the local direction carry the intended delivery beside each sentence.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "language": "de", "seed": 1701, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Bitterness", "label": "Intense Bitterness", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Bitterness", "adapter_sha256": "6aeebbe63751de7d5ef6e4f18717904baa56bf7eb445573349c6ba868b760545", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Bitterness.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Bitterness", "category": "emotion", "head": "emo_Bitterness", "pole": "high", "label": "Intense Bitterness", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 3.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-564f1f7c16dd", "general": "Shift the energy from outright anger to profound, wounded disappointment; the bitterness should feel internal and brittle.", "inline": [ "Begin with a slightly constricted throat sound, suggesting the words are physically difficult to produce.", "Modulate the volume very slightly, a barely perceptible dip in projection that suggests withdrawal.", "Let the final utterance carry a thin, high-frequency edge, like glass about to shatter under pressure." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.37855468960271915, "target_raw": 0.06542543917894364, "ce": 6.2747140312194825, "genuineness": 4.3427893161773685, "identity": 0.5557437974214554, "wer": 0.005631469979296066, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.11886718513980501, "target_raw": 0.17109922394156457, "ce": 6.067648544311523, "genuineness": 4.535463781356811, "identity": 0.5508852855861187, "wer": 0.02144739318652362, "valid_fraction": 1, "identity_delta": -0.0048585118353366855, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.08798827869383732, "target_raw": 0.18366467610001563, "ce": 5.9790480422973635, "genuineness": 4.481364312171936, "identity": 0.5103233557939529, "wer": 0.01768353690550487, "valid_fraction": 0.98, "identity_delta": -0.045420441627502445, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.11886718513980501, "target_raw": 0.17109922394156457, "ce": 6.067648544311523, "genuineness": 4.535463781356811, "identity": 0.5508852855861187, "wer": 0.02144739318652362, "valid_fraction": 1, "identity_delta": -0.0048585118353366855, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.03087890644596771, "ci95_z": [ -0.03822265684648564, 0.11277343922508798 ], "language_deltas": { "en": -0.042187496223688024, "de": -0.03749999908453043, "fr": 0.09804688412608725, "nl": 0.16777343702796102, "es": -0.03173829361599127 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.08860050201416016, "delta_genuineness": -0.05409946918487549, "delta_wer": -0.0037638562810187524, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.03087890644596771, "ci95_z": [ -0.03822265684648564, 0.11277343922508798 ], "language_deltas": { "en": -0.042187496223688024, "de": -0.03749999908453043, "fr": 0.09804688412608725, "nl": 0.16777343702796102, "es": -0.03173829361599127 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.08860050201416016, "delta_genuineness": -0.05409946918487549, "delta_wer": -0.0037638562810187524, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Shift the energy from outright anger to profound, wounded disappointment; the bitterness should feel internal and brittle.\nSCRIPT:\n(Begin with a slightly constricted throat sound, suggesting the words are physically difficult to produce.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Modulate the volume very slightly, a barely perceptible dip in projection that suggests withdrawal.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Let the final utterance carry a thin, high-frequency edge, like glass about to shatter under pressure.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "language": "fr", "seed": 1701, "scale": 0.0, "cfg": 3.0, "frames": 128, "reference": true } }, { "target": "emo_Confusion", "label": "Intense Confusion", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Confusion", "adapter_sha256": "175e1a2d251b18a2b8f840488e3a2dc95a3beb54cc6b352621b6c29d0677ef35", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Confusion.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Confusion", "category": "emotion", "head": "emo_Confusion", "pole": "high", "label": "Intense Confusion", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-de7bf3e68eea", "general": "Employ deliberate, uneven pacing, where pauses are not for breath but for genuine processing of contradictory information.", "inline": [ "Begin with a halting, staggered delivery, allowing significant, uncomfortable gaps between key syllables.", "Mid-phrase, accelerate briefly with a rush of confused energy before immediately decelerating again.", "The final utterance should be delivered with a slow, weighted emphasis on the last syllable, as if confirming a difficult realization." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.382031247119097, "target_raw": 0.11397703737020493, "ce": 6.265818223953247, "genuineness": 4.348563270568848, "identity": 0.591278854906559, "wer": 0.010263504611330698, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.9349218852752208, "target_raw": 1.349846741259098, "ce": 5.695177001953125, "genuineness": 4.35534143447876, "identity": 0.547153008878231, "wer": 0.05000473515805323, "valid_fraction": 0.88, "identity_delta": -0.04412584602832794, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.057070302246604, "target_raw": 1.4150010707974434, "ce": 5.722380638122559, "genuineness": 4.359609141349792, "identity": 0.5283753668516874, "wer": 0.08755632658721904, "valid_fraction": 0.92, "identity_delta": -0.06290348805487156, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.9349218852752208, "target_raw": 1.349846741259098, "ce": 5.695177001953125, "genuineness": 4.35534143447876, "identity": 0.547153008878231, "wer": 0.05000473515805323, "valid_fraction": 0.88, "identity_delta": -0.04412584602832794, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.12214841697138358, "ci95_z": [ -0.08992190036067646, 0.3342187343034436 ], "language_deltas": { "en": 0.4068359100222963, "de": 0.00429686578984041, "fr": -0.057812528459829704, "nl": 0.4265624928413925, "es": -0.16914065533678163 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.027203636169433595, "delta_genuineness": 0.004267706871032715, "delta_wer": 0.0375515914291658, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.12214841697138358, "ci95_z": [ -0.08992190036067646, 0.3342187343034436 ], "language_deltas": { "en": 0.4068359100222963, "de": 0.00429686578984041, "fr": -0.057812528459829704, "nl": 0.4265624928413925, "es": -0.16914065533678163 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.027203636169433595, "delta_genuineness": 0.004267706871032715, "delta_wer": 0.0375515914291658, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Contemplation", "label": "Intense Contemplation", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Contemplation", "adapter_sha256": "1da4275e537a7cedcf7dd03b39b855b9054f94494d3326e893398e70d3d4c42a", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Contemplation.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Contemplation", "category": "emotion", "head": "emo_Contemplation", "pole": "high", "label": "Intense Contemplation", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-27fc7ffa99f2", "general": "Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Establish it from the first phrase. Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Sustain it through the thought. Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.", "Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.", "Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 1.909101578384558, "target_raw": 1.6300405406951903, "ce": 6.253904218673706, "genuineness": 4.341785140037537, "identity": 0.5653184099495411, "wer": 0.010405797101449276, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 3.2307812172662667, "target_raw": 2.408437511920929, "ce": 6.001247835159302, "genuineness": 4.61265905380249, "identity": 0.624268764257431, "wer": 0.03941276115189159, "valid_fraction": 0.94, "identity_delta": 0.05895035430788994, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 3.174374952306793, "target_raw": 2.375217308998108, "ce": 6.026823205947876, "genuineness": 4.58705283164978, "identity": 0.629462673664093, "wer": 0.020836181362497153, "valid_fraction": 0.96, "identity_delta": 0.06414426371455192, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 3.2307812172662667, "target_raw": 2.408437511920929, "ce": 6.001247835159302, "genuineness": 4.61265905380249, "identity": 0.624268764257431, "wer": 0.03941276115189159, "valid_fraction": 0.94, "identity_delta": 0.05895035430788994, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.056406264959474164, "ci95_z": [ -0.14718753890095818, 0.052031278242728206 ], "language_deltas": { "en": 0.1351563254299068, "de": -0.08125003668708257, "fr": -0.07187502466890035, "nl": -0.2000000404822979, "es": -0.0640625483889968 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.02557537078857422, "delta_genuineness": -0.02560622215270996, "delta_wer": -0.018576579789394435, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.056406264959474164, "ci95_z": [ -0.14718753890095818, 0.052031278242728206 ], "language_deltas": { "en": 0.1351563254299068, "de": -0.08125003668708257, "fr": -0.07187502466890035, "nl": -0.2000000404822979, "es": -0.0640625483889968 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.02557537078857422, "delta_genuineness": -0.02560622215270996, "delta_wer": -0.018576579789394435, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Establish it from the first phrase. Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Sustain it through the thought. Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Thoughtful reflection; unhurried pauses, gentle uncertainty and an inward, considered tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Natural clear speech.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Natural clear speech.) [3.2 seconds duration] Te esperaré al otro lado.", "language": "es", "seed": 1701, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "emo_Contempt", "label": "Intense Contempt", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Contempt", "adapter_sha256": "a4b208d7591e4fab030bc3833164f963e75ec2aad8a6dbb625cd14a156e07ced", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Contempt.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Contempt", "category": "emotion", "head": "emo_Contempt", "pole": "high", "label": "Intense Contempt", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-0f7cf1dc94a9", "general": "Focus on precise articulation, exaggerating the crispness of consonants to convey intellectual superiority and disdain.", "inline": [ "Over-articulate the plosives in the initial phrase with sharp, percussive stops.", "Allow vowel sounds to be slightly clipped, as if speaking through a filter of utter dismissal.", "Ensure the final consonants have a distinct, brittle quality that cuts through the air." ], "cfg": 2, "recipe": "P08", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.3217578125825757, "target_raw": 0.06361385107040406, "ce": 6.240560550689697, "genuineness": 4.390236210823059, "identity": 0.5705503010749817, "wer": 0.00857142857142857, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.2915625003303027, "target_raw": 0.07723606124520302, "ce": 6.210060682296753, "genuineness": 4.10178936958313, "identity": 0.5020543923974037, "wer": 0.008689114089571756, "valid_fraction": 1, "identity_delta": -0.06849590867757797, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.3100000013212106, "target_raw": 0.06891822993755341, "ce": 6.2531099700927735, "genuineness": 4.215636763572693, "identity": 0.5579234412312508, "wer": 0.004818865345181135, "valid_fraction": 1, "identity_delta": -0.012626859843730926, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.3374609357659111, "target_raw": 0.05652959764003754, "ce": 6.270107297897339, "genuineness": 4.219904479980468, "identity": 0.5607189834117889, "wer": 0.006766917293233083, "valid_fraction": 1, "identity_delta": -0.00983131766319275, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.02746093444470047, "ci95_z": [ 0.0012890543250093658, 0.07433593510530578 ], "language_deltas": { "en": -0.003710946996201242, "de": 0.11972655630682529, "fr": 0.008984379128783148, "nl": 0.008789056306825277, "es": 0.0035156274772698893 }, "languages_improved": 4, "clear_gain": true, "delta_ce": -0.01699732780456543, "delta_genuineness": -0.0042677164077758786, "delta_wer": -0.0019480519480519478, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.018437500990907955, "ci95_z": [ -0.03585937780757255, -0.0010156241742433703 ], "language_deltas": { "en": -0.02148437252273011, "de": -0.0400390678674181, "fr": 0.0046875, "nl": -0.0388671903901482, "es": 0.00351562582575663 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.043049287796020505, "delta_genuineness": 0.11384739398956299, "delta_wer": -0.0038702487443906205, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Contentment", "label": "Intense Contentment", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Contentment", "adapter_sha256": "4d05cc6756eb76fc9324ed78778689f8376d726a75273011cb350e338682e169", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Contentment.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Contentment", "category": "emotion", "head": "emo_Contentment", "pole": "high", "label": "Intense Contentment", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-e34e2073bf7b", "general": "Treat the delivery as a shared, comfortable secret; intimate, low-key, and deeply resonant.", "inline": [ "Start with a near-whisper quality that immediately builds into a rich, comforting mid-range tone, implying shared tranquility.", "Employ a very slight, almost imperceptible vocal fry on key consonants to enhance the sense of settled, deep comfort.", "Conclude the speech with a sustained, resonant tone that lingers slightly beyond the last word, like a warm echo." ], "cfg": 3, "recipe": "P09", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 3.113105472893091, "target_raw": 1.9226697075366974, "ce": 6.256922426223755, "genuineness": 4.337642979621887, "identity": 0.5603747323155404, "wer": 0.007623376623376623, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 3.859375026101705, "target_raw": 2.298647050857544, "ce": 6.275826005935669, "genuineness": 4.28981954574585, "identity": 0.5565548980236054, "wer": 0.031592885375494074, "valid_fraction": 0.96, "identity_delta": -0.003819834291934967, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 3.8793750190032243, "target_raw": 2.3087232279777528, "ce": 6.253109979629516, "genuineness": 4.206097207069397, "identity": 0.5051676166057587, "wer": 0.01761154368133773, "valid_fraction": 1, "identity_delta": -0.055207115709781644, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 3.868437495452536, "target_raw": 2.3032128047943115, "ce": 6.281227035522461, "genuineness": 4.220783095359803, "identity": 0.5457810742408037, "wer": 0.008365800865800865, "valid_fraction": 1, "identity_delta": -0.014593658074736594, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.010937523550688385, "ci95_z": [ -0.05968745626300729, 0.09624999334517437 ], "language_deltas": { "en": -0.05937500480626299, "de": 0.05312502920729027, "fr": -0.07812488427997692, "nl": -0.023437503142556614, "es": 0.16249998077494818 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.028117055892944335, "delta_genuineness": -0.014685888290405274, "delta_wer": 0.009245742815536865, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.019999992901519263, "ci95_z": [ -0.019687522515493418, 0.06874996347240159 ], "language_deltas": { "en": -0.01718759852839029, "de": 0.1031249759686852, "fr": -0.010937386313395959, "nl": 0.05156248798434264, "es": -0.02656251460364527 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.022716026306152343, "delta_genuineness": -0.08372233867645264, "delta_wer": -0.01398134169415634, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Start with a near-whisper quality that immediately builds into a rich, comforting mid-range tone, implying shared tranquility.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Employ a very slight, almost imperceptible vocal fry on key consonants to enhance the sense of settled, deep comfort.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Conclude the speech with a sustained, resonant tone that lingers slightly beyond the last word, like a warm echo.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "language": "es", "seed": 1702, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Disappointment", "label": "Intense Disappointment", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Disappointment", "adapter_sha256": "97429062fc11c51452edab1d0b2c92279f8580ae4c558345b423c569c468af48", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Disappointment.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Disappointment", "category": "emotion", "head": "emo_Disappointment", "pole": "high", "label": "Intense Disappointment", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-3a61628184eb", "general": "Introduce a hesitant, almost fragile texture to the voice. The energy should feel like it's leaking away rather than being actively suppressed.", "inline": [ "Begin with a slight hesitation, as if the speaker is gathering the courage to utter the words at all.", "Maintain a breathy quality across the middle phrase, suggesting the effort of maintaining composure is draining.", "Let the final inflection trail off into an almost inaudible whisper, signifying complete emotional exhaustion." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.45179687272926294, "target_raw": 0.17289833426475526, "ce": 6.229758520126342, "genuineness": 4.351575784683227, "identity": 0.5865230727195739, "wer": 0.009105966497270844, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.027929688675206014, "target_raw": 0.4701738840341568, "ce": 5.678020820617676, "genuineness": 4.510987215042114, "identity": 0.5520411780476571, "wer": 0.026361283643892338, "valid_fraction": 0.96, "identity_delta": -0.03448189467191696, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.13972656225433547, "target_raw": 0.3917661225795746, "ce": 5.760227479934692, "genuineness": 4.519899196624756, "identity": 0.5416708663105965, "wer": 0.020237674225088414, "valid_fraction": 0.98, "identity_delta": -0.04485220640897751, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.08765624965474174, "target_raw": 0.4282851818203926, "ce": 5.786636810302735, "genuineness": 4.565086679458618, "identity": 0.5679738667607307, "wer": 0.01572557876905703, "valid_fraction": 1, "identity_delta": -0.01854920595884323, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.05207031259959373, "ci95_z": [ -0.1308398406194371, 0.031035150563198027 ], "language_deltas": { "en": 0.039843744821126056, "de": -0.0917968803780614, "fr": -0.1931640521090542, "nl": 0.06904296279097516, "es": -0.08427733812295427 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.02640933036804199, "delta_genuineness": -0.04518748283386231, "delta_wer": 0.004512095456031383, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.11179687357912946, "ci95_z": [ -0.1841406256241207, -0.04076171621699946 ], "language_deltas": { "en": 0.008984380643644697, "de": -0.1072265571551365, "fr": -0.05761718554132331, "nl": -0.17314453747460812, "es": -0.22998046836822403 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.0822066593170166, "delta_genuineness": 0.008911981582641601, "delta_wer": -0.006123609418803927, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Introduce a hesitant, almost fragile texture to the voice. The energy should feel like it's leaking away rather than being actively suppressed.\nSCRIPT:\n(Begin with a slight hesitation, as if the speaker is gathering the courage to utter the words at all.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Maintain a breathy quality across the middle phrase, suggesting the effort of maintaining composure is draining.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Let the final inflection trail off into an almost inaudible whisper, signifying complete emotional exhaustion.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "language": "fr", "seed": 1701, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "emo_Disgust", "label": "Intense Disgust", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Disgust", "adapter_sha256": "c3230a025d7f337dfd18d0e330cc4d57dea737e0a9563621a84d7035f43c89d6", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Disgust.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Disgust", "category": "emotion", "head": "emo_Disgust", "pole": "high", "label": "Intense Disgust", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-981ced9f650c", "general": "A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Establish it from the first phrase. A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Sustain it through the thought. A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.", "A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.", "A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.35927734170791714, "target_raw": 0.046335240453481676, "ce": 6.244690752029419, "genuineness": 4.392244520187378, "identity": 0.5791316756606102, "wer": 0.006453980801806888, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.25701172951130186, "target_raw": 0.3484832213819027, "ce": 6.160974950790405, "genuineness": 4.07178994178772, "identity": 0.5725966927409172, "wer": 0.014838915470494417, "valid_fraction": 0.98, "identity_delta": -0.006534982919692994, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.0770703360361924, "target_raw": 0.7505332718789578, "ce": 5.973746328353882, "genuineness": 3.8139074563980104, "identity": 0.5145791156589985, "wer": 0.030860813099943533, "valid_fraction": 0.98, "identity_delta": -0.0645525600016117, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.9040429842983241, "target_raw": 0.665703169554472, "ce": 6.035282201766968, "genuineness": 3.863425402641296, "identity": 0.5393581134080887, "wer": 0.021000941087897608, "valid_fraction": 1, "identity_delta": -0.039773562252521516, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.1730273517378684, "ci95_z": [ -0.027070313544786572, 0.44271486678279526 ], "language_deltas": { "en": -0.04375000455906875, "de": -0.030273426197308707, "fr": 0.6654297214080739, "nl": 0.26103517424882355, "es": 0.012695293788822042 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.06153587341308594, "delta_genuineness": -0.049517946243286134, "delta_wer": 0.009859872012045927, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.8200586065248906, "ci95_z": [ 0.3551953068581524, 1.2849219061916286 ], "language_deltas": { "en": 0.09433594671311811, "de": 0.36582029074944283, "fr": 1.4298828477378023, "nl": 1.3545898880584495, "es": 0.8556640593656402 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.18722862243652344, "delta_genuineness": -0.2578824853897095, "delta_wer": 0.016021897629449115, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Establish it from the first phrase. A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Sustain it through the thought. A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(A visceral feeling of revulsion; recoiling emphasis, tense consonants and an unmistakably repelled tone. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "language": "es", "seed": 1701, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Distress", "label": "Intense Distress", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Distress", "adapter_sha256": "164fd93e713d030ca27fc0ccfef690072158eaf3b4443e4beeb06976c563c272", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Distress.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Distress", "category": "emotion", "head": "emo_Distress", "pole": "high", "label": "Intense Distress", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 3.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-e66580c21865", "general": "Emphasize vocal constriction and sharp intakes of breath to convey immediate panic. Pitch should fluctuate erratically with rising urgency.", "inline": [ "Start with a tight, constricted vocal delivery suggesting immediate pressure building internally.", "Maintain rapid, shallow breathing patterns throughout the utterance, making the words feel rushed and breathless.", "Conclude with a sudden, sharp vocal break or catch in the breath to signal complete emotional overflow." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.3534374991568529, "target_raw": 0.17004485696554184, "ce": 6.2369068908691405, "genuineness": 4.357977328300476, "identity": 0.579327247440815, "wer": 0.013068699416525503, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.9219335489991143, "target_raw": 2.02990692794323, "ce": 5.950255851745606, "genuineness": 4.088798027038575, "identity": 0.5325792555510998, "wer": 0.030243459439111612, "valid_fraction": 0.94, "identity_delta": -0.04674799188971519, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 3.2826561484691403, "target_raw": 3.1421458983421324, "ce": 5.654858016967774, "genuineness": 3.7308126735687255, "identity": 0.4607018177956343, "wer": 0.09337093722448413, "valid_fraction": 0.84, "identity_delta": -0.1186254296451807, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 3.2185936605694403, "target_raw": 3.0897819590568543, "ce": 5.7452356719970705, "genuineness": 3.8324217557907105, "identity": 0.47871300172060727, "wer": 0.08065979672501411, "valid_fraction": 0.84, "identity_delta": -0.10061424572020769, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.06406248789970001, "ci95_z": [ -0.1278124896315697, 0.2674999611696588 ], "language_deltas": { "en": 0.026562447246338095, "de": -0.2609375007975715, "fr": 0.07812504215735391, "nl": 0.4109374388148674, "es": 0.06562501207751219 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.09037765502929687, "delta_genuineness": -0.10160908222198486, "delta_wer": 0.01271114049947002, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 1.360722599470026, "ci95_z": [ 1.1048241991852847, 1.6404295984767714 ], "language_deltas": { "en": 1.8048826692931776, "de": 1.2388671333648302, "fr": 1.2652343533231443, "nl": 1.5993163307726699, "es": 0.8953125105963078 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.29539783477783205, "delta_genuineness": -0.3579853534698486, "delta_wer": 0.06312747778537252, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Emphasize vocal constriction and sharp intakes of breath to convey immediate panic. Pitch should fluctuate erratically with rising urgency.\nSCRIPT:\n(Start with a tight, constricted vocal delivery suggesting immediate pressure building internally.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Maintain rapid, shallow breathing patterns throughout the utterance, making the words feel rushed and breathless.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Conclude with a sudden, sharp vocal break or catch in the breath to signal complete emotional overflow.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 3.0, "frames": 128, "reference": true } }, { "target": "emo_Elation", "label": "Intense Elation", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Elation", "adapter_sha256": "ab57e90c6bb19d78211bf01dfcf7bed149b0f4d12b5504b7d0617bbaf74a145b", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Elation.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Elation", "category": "emotion", "head": "emo_Elation", "pole": "high", "label": "Intense Elation", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-a8a332cdd043", "general": "Adopt a grand, expansive vocal projection, as if sharing monumental good news with the world.", "inline": [ "Project the initial sounds with full chest resonance, conveying overwhelming, open-hearted gladness.", "Expand the vocal space during the middle, letting the sound swell with expansive, unrestrained glee.", "Deliver the final words with a powerful, sustained upward thrust, a triumphant vocal peak." ], "cfg": 3, "recipe": "P02", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.4042480482425721, "target_raw": 0.1615373659133911, "ce": 6.260099515914917, "genuineness": 4.392872085571289, "identity": 0.5759987241029739, "wer": 0.005714285714285714, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.10560547205550604, "target_raw": 0.3760999935865402, "ce": 6.278049974441529, "genuineness": 4.297601857185364, "identity": 0.5770276230573654, "wer": 0.008492941840767928, "valid_fraction": 1, "identity_delta": 0.0010288989543914796, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.8689257689224069, "target_raw": 1.0762613251805306, "ce": 6.20569221496582, "genuineness": 4.126391482353211, "identity": 0.5354003190994263, "wer": 0.022891329113297075, "valid_fraction": 0.98, "identity_delta": -0.040598405003547666, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.0239257768037704, "target_raw": 1.1876225697994232, "ce": 6.217209081649781, "genuineness": 4.145407881736755, "identity": 0.5207838375866413, "wer": 0.033813288161114244, "valid_fraction": 0.96, "identity_delta": -0.05521488651633263, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.15500000788136342, "ci95_z": [ -0.3934374861039119, 0.062265604104016714 ], "language_deltas": { "en": 0.002929679074200281, "de": -0.11328127540702683, "fr": 0.20937496888935492, "nl": -0.32597656314813844, "es": -0.548046848815207 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.01151686668395996, "delta_genuineness": -0.01901639938354492, "delta_wer": -0.01092195904781717, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.9745312409779129, "ci95_z": [ 0.8117285086834374, 1.0920410059871812 ], "language_deltas": { "en": 1.0597167877686549, "de": 0.6713867101297587, "fr": 1.0101562152597796, "nl": 1.1466308664721159, "es": 0.9847656252592554 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.07235775947570801, "delta_genuineness": -0.17121037483215332, "delta_wer": 0.014398387272529149, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a grand, expansive vocal projection, as if sharing monumental good news with the world.\nSCRIPT:\n(Project the initial sounds with full chest resonance, conveying overwhelming, open-hearted gladness.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Expand the vocal space during the middle, letting the sound swell with expansive, unrestrained glee.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Deliver the final words with a powerful, sustained upward thrust, a triumphant vocal peak.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "language": "nl", "seed": 1701, "scale": 0.0, "cfg": 1.0, "frames": 128, "reference": true } }, { "target": "emo_Fatigue_Exhaustion", "label": "Intense Fatigue Exhaustion", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Fatigue_Exhaustion", "adapter_sha256": "feb003cee765ac7446d2ebb014da10d8bda71fc44cce939de0800ceb8ebae17d", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Fatigue_Exhaustion.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Fatigue_Exhaustion", "category": "emotion", "head": "emo_Fatigue_Exhaustion", "pole": "high", "label": "Intense Fatigue Exhaustion", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-1711f3012766", "general": "Embody exhaustion as a systemic drain, where the voice sounds hollowed out and brittle. Prioritize texture over volume.", "inline": [ "Begin with a breathy onset, implying the lungs are incapable of a full, robust inhalation before speaking.", "Introduce subtle vocal fry or rasping texture on key consonants to signify deep physical strain.", "Deliver the last phrase with a fragile, almost cracking quality, as if the vocal apparatus is on the verge of breaking." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.22037109745475073, "target_raw": 0.40190974444150923, "ce": 6.218479900360108, "genuineness": 4.467933607101441, "identity": 0.6025339433550835, "wer": 0.012038396386222472, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 3.9364063325765475, "target_raw": 2.4115200233459473, "ce": 5.727543363571167, "genuineness": 4.545254383087158, "identity": 0.584985556602478, "wer": 0.03419734613212874, "valid_fraction": 0.98, "identity_delta": -0.01754838675260544, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 4.0684375645113136, "target_raw": 2.482921746969223, "ce": 5.529473075866699, "genuineness": 4.4226205444335935, "identity": 0.5450114373862743, "wer": 0.06336152139241384, "valid_fraction": 0.92, "identity_delta": -0.057522505968809125, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 4.257656324172511, "target_raw": 2.585250153541565, "ce": 5.586183691024781, "genuineness": 4.446595048904419, "identity": 0.5242721079289914, "wer": 0.06323442994838875, "valid_fraction": 0.92, "identity_delta": -0.07826183542609215, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.1892187596611978, "ci95_z": [ -0.510468768891861, 0.12062499579798172 ], "language_deltas": { "en": -0.2867187125434849, "de": 0.3390625018943525, "fr": -0.0062500461533154805, "nl": -0.7531250285875014, "es": -0.23906251291603967 }, "languages_improved": 1, "clear_gain": false, "delta_ce": -0.056710615158081054, "delta_genuineness": -0.023974504470825195, "delta_wer": 0.0001270914440250835, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.13203123193476568, "ci95_z": [ -0.17515627705479803, 0.4303125320662213 ], "language_deltas": { "en": 0.38515628917865385, "de": 0.037499924225900115, "fr": 0.6109375036164911, "nl": -0.43281253289284805, "es": 0.059374975545631294 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.19807028770446777, "delta_genuineness": -0.12263383865356445, "delta_wer": 0.0291641752602851, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Embody exhaustion as a systemic drain, where the voice sounds hollowed out and brittle. Prioritize texture over volume.\nSCRIPT:\n(Begin with a breathy onset, implying the lungs are incapable of a full, robust inhalation before speaking.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Introduce subtle vocal fry or rasping texture on key consonants to signify deep physical strain.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Deliver the last phrase with a fragile, almost cracking quality, as if the vocal apparatus is on the verge of breaking.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "language": "es", "seed": 1702, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Fear", "label": "Intense Fear", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Fear", "adapter_sha256": "e71a1957f3ae2565a2bdca773c39714b5c91808b193dd5011dadbf5e1c40922d", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Fear.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Fear", "category": "emotion", "head": "emo_Fear", "pole": "high", "label": "Intense Fear", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.5, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-e2b016e601f8", "general": "Channel the panic as a physical constriction in the throat, forcing the vocal output to feel strained and barely contained throughout the entire delivery.", "inline": [ "Start with a shallow, rapid intake of breath that immediately tightens the vocal cords.", "Maintain that constricted feeling, allowing only fragmented, urgent bursts of sound.", "End with the sound catching in the throat, a desperate, high-tension release." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.4402783194047616, "target_raw": 0.05241373062133789, "ce": 6.236827487945557, "genuineness": 4.414838299751282, "identity": 0.5837383687496185, "wer": 0.013804536043666478, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.5319091768196555, "target_raw": 0.6471409776806831, "ce": 6.034607019424438, "genuineness": 4.191411280632019, "identity": 0.5505273386836051, "wer": 0.019285714285714285, "valid_fraction": 0.98, "identity_delta": -0.033211030066013336, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.579218738368769, "target_raw": 1.8995647993683815, "ce": 5.739080114364624, "genuineness": 3.7938241243362425, "identity": 0.5089578618109226, "wer": 0.06010178113267358, "valid_fraction": 0.88, "identity_delta": -0.07478050693869591, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 2.3037573387587402, "target_raw": 1.7310536834597587, "ce": 5.871941156387329, "genuineness": 3.9224829149246214, "identity": 0.5097911757975817, "wer": 0.0367287784679089, "valid_fraction": 0.96, "identity_delta": -0.07394719295203686, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.2754613996100286, "ci95_z": [ -0.19946293770941095, 0.7436352500497889 ], "language_deltas": { "en": 0.8392578431766108, "de": 0.7771483992113271, "fr": -0.5250000389737587, "nl": 0.48536376547306936, "es": -0.19946297083710576 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.13286104202270507, "delta_genuineness": -0.1286587905883789, "delta_wer": 0.023373002664764678, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 2.0473095615491133, "ci95_z": [ 1.7786035137181784, 2.3160156093800484 ], "language_deltas": { "en": 2.4427734442747355, "de": 2.0666503837551136, "fr": 1.6164062472596576, "nl": 1.7967773451582316, "es": 2.3139403872978286 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.29552690505981444, "delta_genuineness": -0.39758715629577634, "delta_wer": 0.04081606684695929, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Channel the panic as a physical constriction in the throat, forcing the vocal output to feel strained and barely contained throughout the entire delivery.\nSCRIPT:\n(Start with a shallow, rapid intake of breath that immediately tightens the vocal cords.) [3.2 seconds duration] The road turns left behind the bridge.\n(Maintain that constricted feeling, allowing only fragmented, urgent bursts of sound.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(End with the sound catching in the throat, a desperate, high-tension release.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "language": "en", "seed": 1702, "scale": 0.0, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "emo_Helplessness", "label": "Intense Helplessness", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Helplessness", "adapter_sha256": "22399a2573143f6e3f93de7209548a576c14eeacea1fcdf0d359e7ade3830a79", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Helplessness.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Helplessness", "category": "emotion", "head": "emo_Helplessness", "pole": "high", "label": "Intense Helplessness", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-a0f2e27bb44b", "general": "Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Establish it from the first phrase. Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Sustain it through the thought. Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.", "Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.", "Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.36039062657180443, "target_raw": 0.17579350978136063, "ce": 6.229281921386718, "genuineness": 4.411951313018799, "identity": 0.6204916799068451, "wer": 0.008354978354978355, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.3541015548021043, "target_raw": 1.3897553557157516, "ce": 5.902838144302368, "genuineness": 4.500820074081421, "identity": 0.5970732706785202, "wer": 0.027134481460568415, "valid_fraction": 1, "identity_delta": -0.02341840922832489, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.356425795341983, "target_raw": 1.3914010554552079, "ce": 5.804507827758789, "genuineness": 4.395131511688232, "identity": 0.5969992949068547, "wer": 0.022279810197430333, "valid_fraction": 1, "identity_delta": -0.023492384999990464, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.3541015548021043, "target_raw": 1.3897553557157516, "ce": 5.902838144302368, "genuineness": 4.500820074081421, "identity": 0.5970732706785202, "wer": 0.027134481460568415, "valid_fraction": 1, "identity_delta": -0.02341840922832489, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.00232424053987869, "ci95_z": [ -0.14119140407808198, 0.13208986821654983 ], "language_deltas": { "en": -0.23710937911036725, "de": 0.09228516403503556, "fr": -0.08691404709434367, "nl": 0.21835940804735243, "es": 0.025000056821716375 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.0983303165435791, "delta_genuineness": -0.10568856239318848, "delta_wer": -0.004854671263138082, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.00232424053987869, "ci95_z": [ -0.14119140407808198, 0.13208986821654983 ], "language_deltas": { "en": -0.23710937911036725, "de": 0.09228516403503556, "fr": -0.08691404709434367, "nl": 0.21835940804735243, "es": 0.025000056821716375 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.0983303165435791, "delta_genuineness": -0.10568856239318848, "delta_wer": -0.004854671263138082, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Establish it from the first phrase. Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Sustain it through the thought. Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] The keys are on the kitchen table.\n(Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Unable to change what is happening; fragile breath, pleading uncertainty and exhausted surrender. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The keys are on the kitchen table.\n(Natural clear speech.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Natural clear speech.) [3.2 seconds duration] Take them with you when you leave.", "language": "en", "seed": 1702, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "emo_Hope_Enthusiasm_Optimism", "label": "Intense Hope Enthusiasm Optimism", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Hope_Enthusiasm_Optimism", "adapter_sha256": "3b6baa6acbe4ee3473b5902ece865dc9452e9252af7fc08ebbefe67c95d86fab", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Hope_Enthusiasm_Optimism.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Hope_Enthusiasm_Optimism", "category": "emotion", "head": "emo_Hope_Enthusiasm_Optimism", "pole": "high", "label": "Intense Hope Enthusiasm Optimism", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-d1f4f47e8466", "general": "Channel an eager, energetic anticipation, sounding like someone eagerly awaiting a wonderful outcome.", "inline": [ "Initiate with a slightly higher vocal register, conveying immediate excitement and forward motion.", "Keep the pace lively and conversational, but maintain a focused, positive intent throughout.", "Sustain the energetic quality right to the final word, making the listener feel the momentum." ], "cfg": 3, "recipe": "P07", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.7573437407294099, "target_raw": 1.1373911643028258, "ce": 6.252951135635376, "genuineness": 4.387349238395691, "identity": 0.5753382009267807, "wer": 0.009232072275550536, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.1929882885360084, "target_raw": 1.4678154706954956, "ce": 6.258034420013428, "genuineness": 4.223670082092285, "identity": 0.5115317273139953, "wer": 0.009441558441558442, "valid_fraction": 1, "identity_delta": -0.06380647361278534, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.227578123250253, "target_raw": 1.4940509009361267, "ce": 6.25533390045166, "genuineness": 4.1637967014312744, "identity": 0.5242583581805229, "wer": 0.006453980801806889, "valid_fraction": 1, "identity_delta": -0.051079842746257785, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.4198828122759095, "target_raw": 1.639908673763275, "ce": 6.2472323799133305, "genuineness": 4.168064403533935, "identity": 0.5274669933319092, "wer": 0.007532467532467532, "valid_fraction": 1, "identity_delta": -0.04787120759487152, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.19230468902565667, "ci95_z": [ -0.44640625701126735, -0.021640613991941705 ], "language_deltas": { "en": -0.06640624600934865, "de": -0.2000000314340537, "fr": -0.011523425481999966, "nl": -0.6742187557097011, "es": -0.009374986493179971 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.008101520538330078, "delta_genuineness": -0.004267702102661133, "delta_wer": -0.0010784867306606436, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.03458983471424443, "ci95_z": [ -0.1088867291947625, 0.16343748666508504 ], "language_deltas": { "en": 0.17089840804592327, "de": 0.08945311269037545, "fr": -0.06289062214514941, "nl": -0.21748047729152878, "es": 0.1929687522716016 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.002700519561767578, "delta_genuineness": -0.05987338066101074, "delta_wer": -0.0029875776397515525, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Impatience_and_Irritability", "label": "Intense Impatience and Irritability", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Impatience_and_Irritability", "adapter_sha256": "b8da8d7dee4c0fd9d0be6fc4826fb61eda384028faefcf6508e5a9e1d85427f8", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Impatience_and_Irritability.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Impatience_and_Irritability", "category": "emotion", "head": "emo_Impatience_and_Irritability", "pole": "high", "label": "Intense Impatience and Irritability", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-89f8b6c2e4b7", "general": "Focus on aggressive articulation and a pressurized breath support. The intent is to sound perpetually on the verge of snapping. Maintain forward placement aggressively.", "inline": [ "Initiate with a forced, slightly breathy quality that immediately hardens into sharp consonants. Convey immediate tension.", "Drive the pace forward relentlessly, using vocal fry sparingly to emphasize moments of extreme annoyance. The sound should feel pushed.", "Deliver the final segment with maximum forward projection, as if speaking directly at an unseen irritant." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.5380078112169584, "target_raw": 0.13546290576457978, "ce": 6.235000677108765, "genuineness": 4.364127864837647, "identity": 0.583655651807785, "wer": 0.014050254093732353, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.371542949713679, "target_raw": 1.441850266456604, "ce": 5.900733318328857, "genuineness": 3.881374888420105, "identity": 0.4974919033050537, "wer": 0.05138518727649162, "valid_fraction": 0.96, "identity_delta": -0.08616374850273133, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.36089841453798, "target_raw": 2.1187013357877733, "ce": 5.642675924301147, "genuineness": 3.65160906791687, "identity": 0.42866593196988106, "wer": 0.03203044667003477, "valid_fraction": 1, "identity_delta": -0.15498971983790397, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 2.104296861965387, "target_raw": 1.94315165579319, "ce": 5.717535581588745, "genuineness": 3.7224655294418336, "identity": 0.4748577162623405, "wer": 0.03351200134724162, "valid_fraction": 0.98, "identity_delta": -0.1087979355454445, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.25660155257259354, "ci95_z": [ 0.11937499199545423, 0.41792969643364486 ], "language_deltas": { "en": 0.2148436950708464, "de": 0.3238280713300651, "fr": 0.11992190089906189, "nl": 0.07109373155415047, "es": 0.5533203640088439 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.07485965728759765, "delta_genuineness": -0.07085646152496337, "delta_wer": -0.0014815546772068507, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.9893554648243013, "ci95_z": [ 0.7558202993190452, 1.2149414293770797 ], "language_deltas": { "en": 0.6166015281777872, "de": 1.0856444711393325, "fr": 1.3152344127425226, "nl": 0.7499999869313538, "es": 1.17929692513051 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.25805739402770994, "delta_genuineness": -0.22976582050323485, "delta_wer": -0.019354740606456852, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Interest", "label": "Intense Interest", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Interest", "adapter_sha256": "2527c69758e807c4dfcedc935d5f67569a0894a41bd7104acb461d2bf6e46653", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Interest.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Interest", "category": "emotion", "head": "emo_Interest", "pole": "high", "label": "Intense Interest", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.5, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-bb77d46f4025", "general": "Convey an active, vibrant eagerness; the emotion should feel like the thrill of an unexpected and delightful revelation.", "inline": [ "Start with a quick, energized breath supporting an immediate surge of vocal enthusiasm.", "Use rapid, subtle shifts in vocal color and texture to reflect the speed of mental processing.", "Deliver the final word with an emphatic, open resonance, fully releasing the sense of wonder." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.8261913928305691, "target_raw": 2.19723851442337, "ce": 6.286151475906372, "genuineness": 4.358855981826782, "identity": 0.5689422625303269, "wer": 0.01225108225108225, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.3906250125113875, "target_raw": 2.7131864738464357, "ce": 6.2127612018585205, "genuineness": 4.220155525207519, "identity": 0.5296556118130684, "wer": 0.013907961603613777, "valid_fraction": 0.98, "identity_delta": -0.03928665071725845, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.2596093812811648, "target_raw": 2.5934252738952637, "ce": 6.205453929901123, "genuineness": 4.235468997955322, "identity": 0.5183770287036896, "wer": 0.008203933747412007, "valid_fraction": 1, "identity_delta": -0.05056523382663727, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.2300781017041635, "target_raw": 2.5664307737350462, "ce": 6.241672506332398, "genuineness": 4.2950914478302, "identity": 0.5354235100746155, "wer": 0.0067186147186147184, "valid_fraction": 1, "identity_delta": -0.03351875245571136, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.029531279577001322, "ci95_z": [ -0.07640627179304368, 0.1670313012579723 ], "language_deltas": { "en": -0.1257812436322336, "de": -0.0640625288332462, "fr": 0.01640628092187361, "nl": 0.032031316581365356, "es": 0.28906257284724746 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.036218576431274414, "delta_genuineness": -0.05962244987487793, "delta_wer": 0.0014853190287972893, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.13101563123022264, "ci95_z": [ -0.1989843915918521, -0.06140624799288004 ], "language_deltas": { "en": -0.2238281369968719, "de": -0.11210941241699528, "fr": -0.2113281345516496, "nl": -0.1132812411870113, "es": 0.005468769001414886 }, "languages_improved": 1, "clear_gain": false, "delta_ce": -0.0073072719573974605, "delta_genuineness": 0.015313472747802735, "delta_wer": -0.00570402785620177, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Convey an active, vibrant eagerness; the emotion should feel like the thrill of an unexpected and delightful revelation.\nSCRIPT:\n(Start with a quick, energized breath supporting an immediate surge of vocal enthusiasm.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Use rapid, subtle shifts in vocal color and texture to reflect the speed of mental processing.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Deliver the final word with an emphatic, open resonance, fully releasing the sense of wonder.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "emo_Longing", "label": "Intense Longing", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Longing", "adapter_sha256": "9c33c3ea5517596f44c4baec37ed298d04fb165c488c25e4877ab9381c5f3712", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Longing.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Longing", "category": "emotion", "head": "emo_Longing", "pole": "high", "label": "Intense Longing", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-bd19fff54e5e", "general": "Focus on vocal placement; project the ache from a high, fragile place in the throat, not from the chest.", "inline": [ "Begin with a slightly higher register, almost a fragile whisper carrying significant emotional weight.", "Allow the vowels to stretch and thin, suggesting the yearning is delicate and easily broken.", "Maintain a shimmering, high-frequency texture that fades gently into the silence following the utterance." ], "cfg": 2.5, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.22144531263914963, "target_raw": 0.29429429084062575, "ce": 6.257081289291381, "genuineness": 4.369650764465332, "identity": 0.5816589125990868, "wer": 0.01524334551794506, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.199921884037402, "target_raw": 1.1980436146259308, "ce": 5.812966804504395, "genuineness": 4.489021053314209, "identity": 0.5362272065877914, "wer": 0.010424480172763925, "valid_fraction": 1, "identity_delta": -0.04543170601129532, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.1383007706416721, "target_raw": 1.1588630005717278, "ce": 5.728695058822632, "genuineness": 4.451992473602295, "identity": 0.5308062396198511, "wer": 0.02199513110840342, "valid_fraction": 1, "identity_delta": -0.05085267297923565, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.199921884037402, "target_raw": 1.1980436146259308, "ce": 5.812966804504395, "genuineness": 4.489021053314209, "identity": 0.5362272065877914, "wer": 0.010424480172763925, "valid_fraction": 1, "identity_delta": -0.04543170601129532, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.06162111339572986, "ci95_z": [ -0.19261720909748642, 0.05173823953074554 ], "language_deltas": { "en": 0.11562490801477468, "de": 0.03359382924204925, "fr": -0.29199214828177694, "nl": -0.12177735563264556, "es": -0.04355480032105069 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.0842717456817627, "delta_genuineness": -0.03702857971191406, "delta_wer": 0.011570650935639495, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.06162111339572986, "ci95_z": [ -0.19261720909748642, 0.05173823953074554 ], "language_deltas": { "en": 0.11562490801477468, "de": 0.03359382924204925, "fr": -0.29199214828177694, "nl": -0.12177735563264556, "es": -0.04355480032105069 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.0842717456817627, "delta_genuineness": -0.03702857971191406, "delta_wer": 0.011570650935639495, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Focus on vocal placement; project the ache from a high, fragile place in the throat, not from the chest.\nSCRIPT:\n(Begin with a slightly higher register, almost a fragile whisper carrying significant emotional weight.) [3.2 seconds duration] Escribí el número en el sobre.\n(Allow the vowels to stretch and thin, suggesting the yearning is delicate and easily broken.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Maintain a shimmering, high-frequency texture that fades gently into the silence following the utterance.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "language": "es", "seed": 1702, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "emo_Pain", "label": "Intense Pain", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Pain", "adapter_sha256": "aa00c2b282540639096e5443c4b4d08d5b82c26c7e8cdd5fcc81ffa564a29c8b", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Pain.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Pain", "category": "emotion", "head": "emo_Pain", "pole": "high", "label": "Intense Pain", "floor": 3.0 }, "configuration": { "adapter": true, "scale": 0.75, "cfg": 4.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "manual-abbf6e2776b7", "general": "The performance must convey a sense of desperate, near-breaking vulnerability. The pain is less about the injury and more about the inability to control the reaction to it.", "inline": [ "Begin with an extremely fragile, almost breathy onset, as if the speaker cannot summon the strength to project the initial sound.", "Introduce brief, uncontrolled vocal fry or vocal breaks mid-sentence, punctuating the speech with involuntary physical reactions.", "Conclude with a forced, strained projection on the final word, a momentary surge of desperate will overriding the physical distress." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.34390625048576245, "target_raw": 0.059159489274024965, "ce": 6.221180419921875, "genuineness": 4.423499188423157, "identity": 0.5939027291536331, "wer": 0.010232072275550536, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.14971679607551594, "target_raw": 0.3430774417519569, "ce": 6.0271409034729, "genuineness": 4.489146618843079, "identity": 0.6024945497512817, "wer": 0.015677608051749926, "valid_fraction": 1, "identity_delta": 0.00859182059764862, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.8902636697963414, "target_raw": 0.7690189644694329, "ce": 5.798789148330688, "genuineness": 4.239485726356507, "identity": 0.5498610246554017, "wer": 0.047060982495765104, "valid_fraction": 0.98, "identity_delta": -0.04404170449823141, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.5274609338163012, "target_raw": 0.5603451380133628, "ce": 5.921821041107178, "genuineness": 4.387725777626038, "identity": 0.5914051773771644, "wer": 0.016404103143233578, "valid_fraction": 1, "identity_delta": -0.0024975517764687536, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.3628027359800402, "ci95_z": [ 0.21553710562046097, 0.5911914129008978 ], "language_deltas": { "en": 0.3277343841485265, "de": 0.2131347658982414, "fr": 0.28481445376256326, "nl": 0.8050293048192404, "es": 0.1833007712716293 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.12303189277648925, "delta_genuineness": -0.14824005126953124, "delta_wer": 0.030656879352531526, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.7405468737208255, "ci95_z": [ 0.5923535132103558, 0.9177441543214776 ], "language_deltas": { "en": 0.6820800981121017, "de": 0.504492178472914, "fr": 0.6509277363686502, "nl": 1.0748535251277282, "es": 0.7903808305227332 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.22835175514221193, "delta_genuineness": -0.24966089248657228, "delta_wer": 0.03138337444401518, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. The performance must convey a sense of desperate, near-breaking vulnerability. The pain is less about the injury and more about the inability to control the reaction to it.\nSCRIPT:\n(Begin with an extremely fragile, almost breathy onset, as if the speaker cannot summon the strength to project the initial sound.) [3.2 seconds duration] Your message arrived after I had left.\n(Introduce brief, uncontrolled vocal fry or vocal breaks mid-sentence, punctuating the speech with involuntary physical reactions.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Conclude with a forced, strained projection on the final word, a momentary surge of desperate will overriding the physical distress.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Your message arrived after I had left.\n(Natural clear speech.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Natural clear speech.) [3.2 seconds duration] Let me know when we can talk.", "language": "en", "seed": 1702, "scale": 0.75, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Relief", "label": "Intense Relief", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Relief", "adapter_sha256": "38f77b0677da10399a6065d6756f425a5a4f29860e476a6163885798207c5e1d", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Relief.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Relief", "category": "emotion", "head": "emo_Relief", "pole": "high", "label": "Intense Relief", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-c17f294b7963", "general": "Shift the resonance forward slightly, giving the relief a bright, almost stunned quality without becoming overly bright.", "inline": [ "Project the sound from a relaxed throat space", "Introduce a gentle upward inflection on the key words", "Ensure the final word carries a lingering, warm resonance" ], "cfg": 3, "recipe": "P01", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.02017578290806797, "target_raw": 0.35587691456079484, "ce": 6.244214172363281, "genuineness": 4.4011565208435055, "identity": 0.5936945807933808, "wer": 0.007323546019198193, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.4874121155919209, "target_raw": 0.7101233243942261, "ce": 6.222292404174805, "genuineness": 4.416846647262573, "identity": 0.5549879938364028, "wer": 0.06515222344398545, "valid_fraction": 0.92, "identity_delta": -0.038706586956977845, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.3099951143776253, "target_raw": 0.5863037142157554, "ce": 6.253745384216309, "genuineness": 4.40165855884552, "identity": 0.5622172182798386, "wer": 0.04844760616957413, "valid_fraction": 0.92, "identity_delta": -0.031477362513542176, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.32054687242449, "target_raw": 0.5936678031086922, "ce": 6.2512036895751955, "genuineness": 4.393374238014221, "identity": 0.5607603266835213, "wer": 0.03784679089026915, "valid_fraction": 0.94, "identity_delta": -0.032934254109859465, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.010551758046864737, "ci95_z": [ -0.0899609414967111, 0.08083496699845569 ], "language_deltas": { "en": -0.010839842799196435, "de": -0.11855469787543536, "fr": -0.1009277344667442, "nl": 0.01379394263106715, "es": 0.16376954227598517 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.002541694641113281, "delta_genuineness": 0.008284320831298828, "delta_wer": 0.010600815279304981, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.1774170012142956, "ci95_z": [ -0.24404785919978794, -0.10930665285033217 ], "language_deltas": { "en": -0.217382831749602, "de": -0.10444337127672533, "fr": -0.2951171831963628, "nl": -0.19523926610701003, "es": -0.07490235374177782 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.0314529800415039, "delta_genuineness": -0.015188088417053222, "delta_wer": -0.016704617274411325, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Shift the resonance forward slightly, giving the relief a bright, almost stunned quality without becoming overly bright.\nSCRIPT:\n(Project the sound from a relaxed throat space) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Introduce a gentle upward inflection on the key words) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Ensure the final word carries a lingering, warm resonance) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "emo_Sadness", "label": "Intense Sadness", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Sadness", "adapter_sha256": "13c2e452244411f862e845523c3deb1ea36f0da8613c12ba0e93e589cab985fc", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Sadness.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Sadness", "category": "emotion", "head": "emo_Sadness", "pole": "high", "label": "Intense Sadness", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-dbfdc96d1cab", "general": "Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Establish it from the first phrase. Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Sustain it through the thought. Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.", "Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.", "Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.47003906182676475, "target_raw": 0.13539874106645583, "ce": 6.257557830810547, "genuineness": 4.3779352045059206, "identity": 0.5840153047442436, "wer": 0.005588179936006023, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.5776171821659055, "target_raw": 0.9361255812644959, "ce": 5.878771829605102, "genuineness": 4.646424169540405, "identity": 0.6028088045120239, "wer": 0.03417523056653491, "valid_fraction": 0.96, "identity_delta": 0.018793499767780303, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.4878320364454872, "target_raw": 0.8675025224685669, "ce": 5.77222089767456, "genuineness": 4.531572618484497, "identity": 0.605424752831459, "wer": 0.02024609448522492, "valid_fraction": 0.98, "identity_delta": 0.021409448087215423, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.5776171821659055, "target_raw": 0.9361255812644959, "ce": 5.878771829605102, "genuineness": 4.646424169540405, "identity": 0.6028088045120239, "wer": 0.03417523056653491, "valid_fraction": 0.96, "identity_delta": 0.018793499767780303, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.08978514572041829, "ci95_z": [ -0.15291015425009538, -0.02666013719074121 ], "language_deltas": { "en": -0.17460936808486896, "de": -0.10595700949170767, "fr": -0.15468751279451565, "nl": 0.003125024126800849, "es": -0.016796862357800042 }, "languages_improved": 1, "clear_gain": false, "delta_ce": -0.10655093193054199, "delta_genuineness": -0.1148515510559082, "delta_wer": -0.013929136081309993, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.08978514572041829, "ci95_z": [ -0.15291015425009538, -0.02666013719074121 ], "language_deltas": { "en": -0.17460936808486896, "de": -0.10595700949170767, "fr": -0.15468751279451565, "nl": 0.003125024126800849, "es": -0.016796862357800042 }, "languages_improved": 1, "clear_gain": false, "delta_ce": -0.10655093193054199, "delta_genuineness": -0.1148515510559082, "delta_wer": -0.013929136081309993, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Establish it from the first phrase. Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Sustain it through the thought. Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] I wrote the number on the envelope.\n(Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Deep genuine sadness; fragile breath, lowered energy and restrained grief rather than a neutral reading. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I wrote the number on the envelope.\n(Natural clear speech.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Natural clear speech.) [3.2 seconds duration] Call me when you have a moment.", "language": "en", "seed": 1702, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "emo_Sourness", "label": "Intense Sourness", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Sourness", "adapter_sha256": "9c313f92f0b42fa4a779cac2729cfa69aa4fd593724abb0a7031fd1d8d901623", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Sourness.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Sourness", "category": "emotion", "head": "emo_Sourness", "pole": "high", "label": "Intense Sourness", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.25, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-909f760b761b", "general": "Employ sharp, percussive articulation on consonants, giving the sourness a clipped, impatient edge.", "inline": [ "Attack the first consonant with sudden, sharp precision, immediately conveying annoyance.", "Keep the vowel shapes tight and unrelaxed, resisting any tendency toward openness.", "Finish the utterance with a sudden, abrupt cessation of vocal energy." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.2901562480075331, "target_raw": 0.08964974224567414, "ce": 6.242149085998535, "genuineness": 4.378311710357666, "identity": 0.6032080364227295, "wer": 0.01239462292093871, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.28597656503904906, "target_raw": 0.09145689174532891, "ce": 6.236112642288208, "genuineness": 4.296221151351928, "identity": 0.5841147097945213, "wer": 0.011096837944664031, "valid_fraction": 1, "identity_delta": -0.01909332662820816, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.06460937308830876, "target_raw": 0.2430380541086197, "ce": 6.174874639511108, "genuineness": 4.208607630729675, "identity": 0.5608653366565705, "wer": 0.008484848484848484, "valid_fraction": 1, "identity_delta": -0.04234269976615906, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.011601560860253556, "target_raw": 0.21008709773421289, "ce": 6.198464374542237, "genuineness": 4.225427384376526, "identity": 0.5334365478157997, "wer": 0.00823640127987954, "valid_fraction": 1, "identity_delta": -0.06977148860692978, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.0762109339485623, "ci95_z": [ 0.01593749149162768, 0.15835937627625585 ], "language_deltas": { "en": 0.04726562540387843, "de": 0.021093729159872877, "fr": 0.08515625145396234, "nl": 0.2324218768578408, "es": -0.004882813132742886 }, "languages_improved": 4, "clear_gain": true, "delta_ce": -0.02358973503112793, "delta_genuineness": -0.016819753646850587, "delta_wer": 0.000248447204968944, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.35058593812735783, "ci95_z": [ 0.19316405989094534, 0.5053906289795489 ], "language_deltas": { "en": 0.1390624964458698, "de": 0.4804687474690285, "fr": 0.5269531397281002, "nl": 0.49628905898625764, "es": 0.11015624800753307 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.06123800277709961, "delta_genuineness": -0.08761352062225342, "delta_wer": -0.0026119894598155466, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Employ sharp, percussive articulation on consonants, giving the sourness a clipped, impatient edge.\nSCRIPT:\n(Attack the first consonant with sudden, sharp precision, immediately conveying annoyance.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Keep the vowel shapes tight and unrelaxed, resisting any tendency toward openness.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Finish the utterance with a sudden, abrupt cessation of vocal energy.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "language": "de", "seed": 1702, "scale": 0.0, "cfg": 1.25, "frames": 128, "reference": true } }, { "target": "emo_Teasing", "label": "Intense Teasing", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Teasing", "adapter_sha256": "15a39dab0aa8f8651831c3a629ee387b98c46daa11031c96fe1473104fa08596", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Teasing.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Teasing", "category": "emotion", "head": "emo_Teasing", "pole": "high", "label": "Intense Teasing", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.5, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-c46a333eb8dd", "general": "Focus on a deliberate, exaggerated sweetness that hints at underlying, playful challenge.", "inline": [ "Begin with an overly sweet, almost saccharine inflection, immediately setting up the ironic contrast.", "Introduce subtle, rhythmic vocal emphasis on the unstressed syllables to create a buoyant, teasing cadence.", "End with a clear, bright resonance, ensuring the underlying flirtatious challenge is perfectly audible." ], "cfg": 2.5, "recipe": "P09", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.3035058635922404, "target_raw": 0.17814444601535798, "ce": 6.261211471557617, "genuineness": 4.350320591926574, "identity": 0.5787997949123382, "wer": 0.016744588744588743, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 2.1701757692842025, "target_raw": 1.5090199303627014, "ce": 6.255175065994263, "genuineness": 4.159277949333191, "identity": 0.5305546429753304, "wer": 0.012233320455288419, "valid_fraction": 1, "identity_delta": -0.0482451519370079, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.0236718814567705, "target_raw": 1.4301987797021867, "ce": 6.257716703414917, "genuineness": 4.074551415443421, "identity": 0.5226720720529556, "wer": 0.02474120082815735, "valid_fraction": 0.98, "identity_delta": -0.05612772285938263, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 2.1701757692842025, "target_raw": 1.5090199303627014, "ce": 6.255175065994263, "genuineness": 4.159277949333191, "identity": 0.5305546429753304, "wer": 0.012233320455288419, "valid_fraction": 1, "identity_delta": -0.0482451519370079, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.14650388782743207, "ci95_z": [ -0.3213476036567492, 0.08248047444944137 ], "language_deltas": { "en": -0.32734372524616195, "de": -0.15791018833911014, "fr": 0.29921876194415964, "nl": -0.3738280348994913, "es": -0.17265625259655654 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.002541637420654297, "delta_genuineness": -0.08472653388977051, "delta_wer": 0.012507880372868931, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.14650388782743207, "ci95_z": [ -0.3213476036567492, 0.08248047444944137 ], "language_deltas": { "en": -0.32734372524616195, "de": -0.15791018833911014, "fr": 0.29921876194415964, "nl": -0.3738280348994913, "es": -0.17265625259655654 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.002541637420654297, "delta_genuineness": -0.08472653388977051, "delta_wer": 0.012507880372868931, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Begin with an overly sweet, almost saccharine inflection, immediately setting up the ironic contrast.) [3.2 seconds duration] The road turns left behind the bridge.\n(Introduce subtle, rhythmic vocal emphasis on the unstressed syllables to create a buoyant, teasing cadence.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(End with a clear, bright resonance, ensuring the underlying flirtatious challenge is perfectly audible.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "language": "en", "seed": 1702, "scale": 0.0, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "emo_Thankfulness_Gratitude", "label": "Intense Thankfulness Gratitude", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Thankfulness_Gratitude", "adapter_sha256": "8050b7492421efc0c6d7e488ddc13e015cf77cc9797eadfad5dac5ebdd2b9f70", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Thankfulness_Gratitude.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Thankfulness_Gratitude", "category": "emotion", "head": "emo_Thankfulness_Gratitude", "pole": "high", "label": "Intense Thankfulness Gratitude", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 3.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-4df4189531b0", "general": "Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Establish it from the first phrase. Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Sustain it through the thought. Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.", "Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.", "Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 2, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.2584472665966884, "target_raw": 0.1402481535077095, "ce": 6.250886039733887, "genuineness": 4.370529451370239, "identity": 0.5567031341791153, "wer": 0.008362507058159233, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.18353027180987982, "target_raw": 0.19203736364841462, "ce": 6.267406759262085, "genuineness": 4.588433547019958, "identity": 0.6213389438390732, "wer": 0.007489177489177489, "valid_fraction": 1, "identity_delta": 0.06463580965995788, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.14766356903318414, "target_raw": 0.2168315804004669, "ce": 6.282021322250366, "genuineness": 4.5880569648742675, "identity": 0.6553041976690293, "wer": 0.011998870694522868, "valid_fraction": 1, "identity_delta": 0.09860106348991395, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.18353027180987982, "target_raw": 0.19203736364841462, "ce": 6.267406759262085, "genuineness": 4.588433547019958, "identity": 0.6213389438390732, "wer": 0.007489177489177489, "valid_fraction": 1, "identity_delta": 0.06463580965995788, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.03586670277669568, "ci95_z": [ -0.009672851586918515, 0.09512207785108423 ], "language_deltas": { "en": -0.026782227812896253, "de": 0.04880371733262424, "fr": -0.0009155267606527608, "nl": 0.007031251212505451, "es": 0.15119629991189773 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.01461456298828125, "delta_genuineness": -0.00037658214569091797, "delta_wer": 0.004509693205345379, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.03586670277669568, "ci95_z": [ -0.009672851586918515, 0.09512207785108423 ], "language_deltas": { "en": -0.026782227812896253, "de": 0.04880371733262424, "fr": -0.0009155267606527608, "nl": 0.007031251212505451, "es": 0.15119629991189773 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.01461456298828125, "delta_genuineness": -0.00037658214569091797, "delta_wer": 0.004509693205345379, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Establish it from the first phrase. Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Sustain it through the thought. Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Specific heartfelt gratitude; warm emphasis and the vulnerable relief of feeling genuinely helped. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Natural clear speech.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Natural clear speech.) [3.2 seconds duration] Neem ze mee als je weggaat.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 3.0, "frames": 128, "reference": true } }, { "target": "vn_S_ASMR_reg", "label": "Strong ASMR Style", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_ASMR_reg", "adapter_sha256": "bbad53ee1ff6299eaa954816f538c013931c1c71fec0dbe97985af785ce490a8", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_ASMR_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_ASMR_reg", "category": "talking_style", "head": "vn_S_ASMR_reg", "pole": "high", "label": "Strong ASMR Style", "floor": 4.0, "canonical_dimension_name": "ASMR Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-763a50e7d58f", "general": "Very close, gentle ASMR speech; breath detail and soft consonants, without mouth sound effects or added words. Establish it immediately. Very close, gentle ASMR speech; breath detail and soft consonants, without mouth sound effects or added words. Sustain the delivery. Very close, gentle ASMR speech; breath detail and soft consonants, without mouth sound effects or added words. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Very close, gentle ASMR speech; breath detail and soft consonants, without mouth sound effects or added words. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.", "Very close, gentle ASMR speech; breath detail and soft consonants, without mouth sound effects or added words. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.", "Very close, gentle ASMR speech; breath detail and soft consonants, without mouth sound effects or added words. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.03717772730589798, "target_raw": 1.546500015258789, "ce": 6.277096824645996, "genuineness": 4.413583059310913, "identity": 0.5892024129629135, "wer": 0.007492941840767927, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.3238281180018254, "target_raw": 3.422941837310791, "ce": 5.897397422790528, "genuineness": 4.447975788116455, "identity": 0.5337682750821113, "wer": 0.01115050472029877, "valid_fraction": 1, "identity_delta": -0.055434137880802155, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.3501562403268474, "target_raw": 3.4592408657073976, "ce": 5.709553251266479, "genuineness": 4.376679944992065, "identity": 0.4986606428027153, "wer": 0.04616327875023527, "valid_fraction": 0.94, "identity_delta": -0.09054177016019821, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.3375000060524753, "target_raw": 3.441791501045227, "ce": 5.7461292171478275, "genuineness": 4.370278406143188, "identity": 0.521460379064083, "wer": 0.0458408868019852, "valid_fraction": 0.96, "identity_delta": -0.06774203389883042, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.012656234274372201, "ci95_z": [ -0.0596875228048623, 0.09015622303406093 ], "language_deltas": { "en": -0.04648442836445862, "de": -0.09765625675499474, "fr": 0.13984372162902203, "nl": 0.07773437864769714, "es": -0.010156243785404806 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.036575965881347657, "delta_genuineness": 0.006401538848876953, "delta_wer": 0.00032239194825007216, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.026328122325022088, "ci95_z": [ -0.019296868380105144, 0.0690624957848833 ], "language_deltas": { "en": -0.012499987030410043, "de": 0.07421875513379607, "fr": 0.0484374583892323, "nl": -0.0527343700013039, "es": 0.07421875513379601 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.18784417152404786, "delta_genuineness": -0.07129584312438965, "delta_wer": 0.0350127740299365, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "vn_S_AUTH_reg", "label": "Strong Authoritative Style", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_AUTH_reg", "adapter_sha256": "a5564f04252aee3f2cea036f19455d8fa56ba9b8547380203c76690c0a4690e5", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_AUTH_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_AUTH_reg", "category": "talking_style", "head": "vn_S_AUTH_reg", "pole": "high", "label": "Strong Authoritative Style", "floor": 4.0, "canonical_dimension_name": "Authoritative Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-553924db97fc", "general": "Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Establish it immediately. Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Sustain the delivery. Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Preserve the intended style through the end. Let the local direction carry the intended delivery beside each sentence.", "inline": [ "Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Establish it immediately. Let the local direction carry the intended delivery beside each sentence.", "Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Sustain the delivery. Let the local direction carry the intended delivery beside each sentence.", "Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Preserve the intended style through the end. Let the local direction carry the intended delivery beside each sentence." ], "cfg": 3, "recipe": "P09", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 1.555937495732801, "target_raw": 2.827916007041931, "ce": 6.249774055480957, "genuineness": 4.314170608520508, "identity": 0.5820743036270142, "wer": 0.009243459439111612, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 2.308749980062499, "target_raw": 3.6099956798553468, "ce": 6.2729666137695315, "genuineness": 4.106935706138611, "identity": 0.5473741683363914, "wer": 0.005274327122153209, "valid_fraction": 1, "identity_delta": -0.03470013529062271, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.3879687283592053, "target_raw": 3.692294225692749, "ce": 6.257081327438354, "genuineness": 4.0835889005661015, "identity": 0.5499366283416748, "wer": 0.0019047619047619048, "valid_fraction": 1, "identity_delta": -0.032137675285339354, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 2.308749980062499, "target_raw": 3.6099956798553468, "ce": 6.2729666137695315, "genuineness": 4.106935706138611, "identity": 0.5473741683363914, "wer": 0.005274327122153209, "valid_fraction": 1, "identity_delta": -0.03470013529062271, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.07921874829670629, "ci95_z": [ -0.017187450263823296, 0.19531244847088255 ], "language_deltas": { "en": 0.12187494441883648, "de": -0.06874992039343022, "fr": 0.3054687045489517, "nl": 0.014062576558570461, "es": 0.023437436350603046 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.015885286331176758, "delta_genuineness": -0.023346805572509767, "delta_wer": -0.0033695652173913043, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.07921874829670629, "ci95_z": [ -0.017187450263823296, 0.19531244847088255 ], "language_deltas": { "en": 0.12187494441883648, "de": -0.06874992039343022, "fr": 0.3054687045489517, "nl": 0.014062576558570461, "es": 0.023437436350603046 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.015885286331176758, "delta_genuineness": -0.023346805572509767, "delta_wer": -0.0033695652173913043, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Establish it immediately. Let the local direction carry the intended delivery beside each sentence.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Sustain the delivery. Let the local direction carry the intended delivery beside each sentence.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Calm, unmistakable authority; grounded resonance, firm attacks and decisive sentence endings. Preserve the intended style through the end. Let the local direction carry the intended delivery beside each sentence.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Natural clear speech.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Natural clear speech.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "language": "de", "seed": 1701, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_S_CASU_reg", "label": "Strong Casual Style", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_CASU_reg", "adapter_sha256": "a1f50ff5bbaf736e1a2e421bb53b5180206e1d541b2119d78bdb185de1f8daa2", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_CASU_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_CASU_reg", "category": "talking_style", "head": "vn_S_CASU_reg", "pole": "high", "label": "Strong Casual Style", "floor": 4.0, "canonical_dimension_name": "Casual Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-e8c5567cec7f", "general": "Emphasize the casual, unscripted nature of the speech; the speaker is rambling slightly, thinking out loud.", "inline": [ "Allow the pace to vary naturally, speeding up slightly on familiar phrases and slowing down for emphasis.", "Incorporate subtle vocal texture shifts that suggest momentary distraction or shifting focus.", "End the clip with a sense of lingering thought, as if the speaker is about to change the subject abruptly." ], "cfg": 3, "recipe": "P04", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.7763769473138078, "target_raw": 2.7598952579498293, "ce": 6.239607410430908, "genuineness": 4.378813800811767, "identity": 0.5984692323207855, "wer": 0.010144927536231883, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.030966782904963304, "target_raw": 3.7899839735031127, "ce": 5.987586421966553, "genuineness": 4.533329925537109, "identity": 0.6220511162281036, "wer": 0.027190055177469363, "valid_fraction": 0.98, "identity_delta": 0.023581883907318114, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.07578857654381994, "target_raw": 3.8471720361709596, "ce": 5.962408142089844, "genuineness": 4.535338220596313, "identity": 0.579516797363758, "wer": 0.027015433841520796, "valid_fraction": 1, "identity_delta": -0.018952434957027434, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.030966782904963304, "target_raw": 3.7899839735031127, "ce": 5.987586421966553, "genuineness": 4.533329925537109, "identity": 0.6220511162281036, "wer": 0.027190055177469363, "valid_fraction": 0.98, "identity_delta": 0.023581883907318114, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.04482179363885664, "ci95_z": [ -0.03737792713136011, 0.14386721944558975 ], "language_deltas": { "en": 0.2255127411318663, "de": -0.014770457627212083, "fr": 0.05428469534700182, "nl": 0.057568331459561896, "es": -0.09848634211693473 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.025178279876708985, "delta_genuineness": 0.0020082950592041015, "delta_wer": -0.00017462133594856697, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.04482179363885664, "ci95_z": [ -0.03737792713136011, 0.14386721944558975 ], "language_deltas": { "en": 0.2255127411318663, "de": -0.014770457627212083, "fr": 0.05428469534700182, "nl": 0.057568331459561896, "es": -0.09848634211693473 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.025178279876708985, "delta_genuineness": 0.0020082950592041015, "delta_wer": -0.00017462133594856697, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Emphasize the casual, unscripted nature of the speech; the speaker is rambling slightly, thinking out loud.\nSCRIPT:\n(Allow the pace to vary naturally, speeding up slightly on familiar phrases and slowing down for emphasis.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Incorporate subtle vocal texture shifts that suggest momentary distraction or shifting focus.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(End the clip with a sense of lingering thought, as if the speaker is about to change the subject abruptly.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "language": "de", "seed": 1701, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "vn_S_CONV_reg", "label": "Strong Conversational Style", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_CONV_reg", "adapter_sha256": "ae99a479de02d4ad74a01ea9278c99e563ccd2531a90d0df85d925aea216f773", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_CONV_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_CONV_reg", "category": "talking_style", "head": "vn_S_CONV_reg", "pole": "high", "label": "Strong Conversational Style", "floor": 4.0, "canonical_dimension_name": "Conversational Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-ea97be76622a", "general": "Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.", "inline": [ "Employ slight vocal breathiness during transitions to mimic natural thought retrieval.", "Vary the projection level slightly, moving between confidential softness and moderate conversational volume.", "Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.26108397268219125, "target_raw": 2.874166555404663, "ce": 6.240560531616211, "genuineness": 4.325969552993774, "identity": 0.5837892112135887, "wer": 0.010263504611330698, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.7879101627906956, "target_raw": 3.5618593454360963, "ce": 6.070905027389526, "genuineness": 4.5003179264068605, "identity": 0.5285264265537262, "wer": 0.02275270686597918, "valid_fraction": 1, "identity_delta": -0.05526278465986252, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.6863623100960149, "target_raw": 3.4293038129806517, "ce": 6.172174110412597, "genuineness": 4.479983587265014, "identity": 0.5721769249439239, "wer": 0.015780726519856954, "valid_fraction": 1, "identity_delta": -0.011612286269664764, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.625039065949606, "target_raw": 3.3492554903030394, "ce": 6.174477519989014, "genuineness": 4.490652866363526, "identity": 0.5690372335910797, "wer": 0.020416619612271786, "valid_fraction": 0.98, "identity_delta": -0.014751977622509002, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.061323244146408884, "ci95_z": [ -0.03576171592646003, 0.13406738749595548 ], "language_deltas": { "en": 0.08447266059322164, "de": -0.11611327157362329, "fr": 0.08984374807364297, "nl": 0.07968748701492676, "es": 0.16872559662387634 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.002303409576416016, "delta_genuineness": -0.010669279098510741, "delta_wer": -0.004635893092414831, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.10154785269468067, "ci95_z": [ -0.11517577280078406, -0.08536134507961728 ], "language_deltas": { "en": -0.11953127618418616, "de": -0.10712889311331532, "fr": -0.09433595556851537, "nl": -0.11484370926111635, "es": -0.07189942934627015 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.10126908302307129, "delta_genuineness": -0.020334339141845702, "delta_wer": -0.006971980346122223, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "vn_S_DRAM_reg", "label": "Strong Dramatic Style", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_DRAM_reg", "adapter_sha256": "3e9c0e51f6571be4e2970f4775c0bb856eabc35de792219325537ef066210d2f", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_DRAM_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_DRAM_reg", "category": "talking_style", "head": "vn_S_DRAM_reg", "pole": "high", "label": "Strong Dramatic Style", "floor": 4.0, "canonical_dimension_name": "Dramatic Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-21a285777a88", "general": "Employ extreme dynamic shifts. Start almost whispered, suggesting secrecy or fear, then transition abruptly to a full, resonant declaration.", "inline": [ "Initiate at the very edge of audibility, focusing on breath control for intimate delivery.", "Execute a sudden, powerful surge in volume and projection, making the shift jarring and intentional.", "Hold the peak energy level briefly, then allow a controlled, slow dissipation of power." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.3134814414265855, "target_raw": 2.4488737869262693, "ce": 6.255175065994263, "genuineness": 4.342161731719971, "identity": 0.5841461846232414, "wer": 0.00927159796725014, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.992753914461204, "target_raw": 3.9370853090286255, "ce": 5.76761414527893, "genuineness": 4.151809430122375, "identity": 0.46429080963134767, "wer": 0.015734236777715037, "valid_fraction": 1, "identity_delta": -0.11985537499189376, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.0135937372806039, "target_raw": 3.960828399658203, "ce": 5.682071561813355, "genuineness": 4.116600837707519, "identity": 0.4543541517108679, "wer": 0.02412020168999574, "valid_fraction": 0.98, "identity_delta": -0.12979203291237354, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.992753914461204, "target_raw": 3.9370853090286255, "ce": 5.76761414527893, "genuineness": 4.151809430122375, "identity": 0.46429080963134767, "wer": 0.015734236777715037, "valid_fraction": 1, "identity_delta": -0.11985537499189376, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.020839822819399806, "ci95_z": [ -0.09390625801093073, 0.14990232131530667 ], "language_deltas": { "en": -0.17109374722069745, "de": 0.24462888607962077, "fr": -0.013085976369362973, "nl": 0.03593747743860317, "es": 0.0078124741688355035 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.08554258346557617, "delta_genuineness": -0.03520859241485596, "delta_wer": 0.008385964912280702, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.020839822819399806, "ci95_z": [ -0.09390625801093073, 0.14990232131530667 ], "language_deltas": { "en": -0.17109374722069745, "de": 0.24462888607962077, "fr": -0.013085976369362973, "nl": 0.03593747743860317, "es": 0.0078124741688355035 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.08554258346557617, "delta_genuineness": -0.03520859241485596, "delta_wer": 0.008385964912280702, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "vn_S_FORM_reg", "label": "Strong Formal Style", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_FORM_reg", "adapter_sha256": "dba292228ce453db159f6bc05a0aef55c451d9a91b7103e56beab005fcbe293e", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_FORM_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_FORM_reg", "category": "talking_style", "head": "vn_S_FORM_reg", "pole": "high", "label": "Strong Formal Style", "floor": 4.0, "canonical_dimension_name": "Formal Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-b9f763b6bfff", "general": "Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Establish it immediately. Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Sustain the delivery. Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.", "Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.", "Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 1.2911230669522638, "target_raw": 2.5710305404663085, "ce": 6.240878276824951, "genuineness": 4.323082594871521, "identity": 0.587313680499792, "wer": 0.004679089026915113, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.5998828047507996, "target_raw": 2.882364046573639, "ce": 6.228964214324951, "genuineness": 4.279401307106018, "identity": 0.6227012559771538, "wer": 0.005631469979296066, "valid_fraction": 1, "identity_delta": 0.03538757547736168, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.549218755874245, "target_raw": 2.8312776708602905, "ce": 6.257557849884034, "genuineness": 4.280154490470887, "identity": 0.621495686173439, "wer": 0.016623376623376623, "valid_fraction": 0.98, "identity_delta": 0.034182005673646924, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.5218750152213139, "target_raw": 2.803705997467041, "ce": 6.2515214157104495, "genuineness": 4.290949263572693, "identity": 0.6125825208425522, "wer": 0.006680337206652996, "valid_fraction": 1, "identity_delta": 0.025268840342760085, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.02734374065293109, "ci95_z": [ -0.08320312716127878, 0.13789060846714096 ], "language_deltas": { "en": -0.08515622210657309, "de": 0.17109376163765488, "fr": -0.008984393749555198, "nl": -0.11835939892184628, "es": 0.17812495640497514 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.006036434173583984, "delta_genuineness": -0.01079477310180664, "delta_wer": 0.009943039416723627, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.05066404887655473, "ci95_z": [ -0.17238281007087045, 0.04582034901083332 ], "language_deltas": { "en": -0.2859374922415634, "de": 0.0777344395612762, "fr": -0.04531250923623407, "nl": -0.04101561807282441, "es": 0.041210935606572054 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.02859363555908203, "delta_genuineness": 0.000753183364868164, "delta_wer": 0.010991906644080555, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Establish it immediately. Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Sustain the delivery. Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Formal, composed public speech; precise articulation, measured rhythm and professional restraint. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "language": "nl", "seed": 1701, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "vn_S_MONO_reg", "label": "Strong introspective monologue delivery, not monotone speech", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_MONO_reg", "adapter_sha256": "dfab837fc53f0a7fb5f8620d41652318200528549d6cc3cae20622126cbccf00", "recipe_status": "corrected_taxonomy_new_search", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_MONO_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_MONO_reg", "category": "talking_style", "head": "vn_S_MONO_reg", "pole": "high", "label": "Strong introspective monologue delivery, not monotone speech", "floor": 4.0, "canonical_dimension_name": "Monologue Style" }, "configuration": { "adapter": true, "scale": 1.0, "cfg": 5.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "canonical-extension-P02", "general": "An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "inline": [ "An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality." ], "recipe": "P02", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 0.4482177785619005, "target_raw": 3.2363388919830323, "ce": 6.212602329254151, "genuineness": 4.331115899085998, "identity": 0.6053401204943657, "wer": 0.009232072275550536, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.0026562520349374, "target_raw": 3.9417580366134644, "ce": 6.221180410385132, "genuineness": 4.490401792526245, "identity": 0.6051376783847808, "wer": 0.008488612836438923, "valid_fraction": 1, "identity_delta": -0.00020244210958480835, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.0978125094573348, "target_raw": 4.062826557159424, "ce": 6.143421659469604, "genuineness": 4.416093521118164, "identity": 0.6289109975099564, "wer": 0.009318181818181817, "valid_fraction": 1, "identity_delta": 0.023570877015590668, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.0026562520349374, "target_raw": 3.9417580366134644, "ce": 6.221180410385132, "genuineness": 4.490401792526245, "identity": 0.6051376783847808, "wer": 0.008488612836438923, "valid_fraction": 1, "identity_delta": -0.00020244210958480835, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.09515625742239739, "ci95_z": [ 0.05718748913724092, 0.13312502570755386 ], "language_deltas": { "en": 0.04140623762933772, "de": 0.12578130116623049, "fr": 0.04726562042505095, "nl": 0.15273439831394053, "es": 0.10859372957742729 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.07775875091552735, "delta_genuineness": -0.07430827140808105, "delta_wer": 0.0008295689817428949, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.09515625742239739, "ci95_z": [ 0.05718748913724092, 0.13312502570755386 ], "language_deltas": { "en": 0.04140623762933772, "de": 0.12578130116623049, "fr": 0.04726562042505095, "nl": 0.15273439831394053, "es": 0.10859372957742729 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.07775875091552735, "delta_genuineness": -0.07430827140808105, "delta_wer": 0.0008295689817428949, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality.\nSCRIPT:\n(An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(An inward, connected spoken monologue, thinking aloud with natural expressive pitch and one continuous personal train of thought. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "language": "nl", "seed": 1701, "scale": 1.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_S_NARR_reg", "label": "Strong Narrator Style", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_NARR_reg", "adapter_sha256": "d92ba3b17d453de4df907cc11428cfc0608447da31df0c623ea19515d7223b7d", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_NARR_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_NARR_reg", "category": "talking_style", "head": "vn_S_NARR_reg", "pole": "high", "label": "Strong Narrator Style", "floor": 4.0, "canonical_dimension_name": "Narrator Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-9ea232963c9d", "general": "Focus on precise articulation, ensuring every consonant is cleanly formed while maintaining a relaxed jaw posture.", "inline": [ "Deliver the opening segment with crisp, clear articulation, emphasizing the initial consonant sounds.", "Mid-section demands meticulous enunciation of all mid-range vowels and connecting sounds.", "End with sharp, defined final consonants, giving the narration a clean, authoritative stop." ], "cfg": 2.5, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 1.5735937465219048, "target_raw": 3.6661133670806887, "ce": 6.230711612701416, "genuineness": 4.349567494392395, "identity": 0.5985677176713944, "wer": 0.012680342159747194, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.8620312308445206, "target_raw": 4.0800893592834475, "ce": 6.261052646636963, "genuineness": 4.125387306213379, "identity": 0.5303111308813095, "wer": 0.010303030303030302, "valid_fraction": 1, "identity_delta": -0.06825658679008484, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.886874988579389, "target_raw": 4.115746026039123, "ce": 6.237542333602906, "genuineness": 4.102291498184204, "identity": 0.5742505073547364, "wer": 0.006540560888386975, "valid_fraction": 1, "identity_delta": -0.02431721031665802, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.8620312308445206, "target_raw": 4.0800893592834475, "ce": 6.261052646636963, "genuineness": 4.125387306213379, "identity": 0.5303111308813095, "wer": 0.010303030303030302, "valid_fraction": 1, "identity_delta": -0.06825658679008484, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.024843757734868396, "ci95_z": [ -0.017812472071414794, 0.06749998754115158 ], "language_deltas": { "en": -0.009374934591045792, "de": 0.014062551392749766, "fr": 0.08281245379843705, "nl": 0.07890623935806702, "es": -0.04218752128386607 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.023510313034057616, "delta_genuineness": -0.023095808029174804, "delta_wer": -0.0037624694146433275, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.024843757734868396, "ci95_z": [ -0.017812472071414794, 0.06749998754115158 ], "language_deltas": { "en": -0.009374934591045792, "de": 0.014062551392749766, "fr": 0.08281245379843705, "nl": 0.07890623935806702, "es": -0.04218752128386607 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.023510313034057616, "delta_genuineness": -0.023095808029174804, "delta_wer": -0.0037624694146433275, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "vn_S_PLAY_reg", "label": "Strong Playful Style", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_PLAY_reg", "adapter_sha256": "911549d120d62d131404c1fb043060338b69f0920da5972d99417d7cc1c22438", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_PLAY_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_PLAY_reg", "category": "talking_style", "head": "vn_S_PLAY_reg", "pole": "high", "label": "Strong Playful Style", "floor": 4.0, "canonical_dimension_name": "Playful Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-3fb618b13504", "general": "Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Establish it immediately. Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Sustain the delivery. Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.", "Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.", "Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.3973535129769969, "target_raw": 1.752017261981964, "ce": 6.250091753005981, "genuineness": 4.30099091053009, "identity": 0.5770703337341547, "wer": 0.007453416149068323, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.05236571680928542, "target_raw": 2.200501010417938, "ce": 6.268836431503296, "genuineness": 4.264840903282166, "identity": 0.5568684732913971, "wer": 0.006421513269339356, "valid_fraction": 1, "identity_delta": -0.020201860442757607, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.45795899765441767, "target_raw": 2.863922567367554, "ce": 6.2791619300842285, "genuineness": 4.227059121131897, "identity": 0.5882936984300613, "wer": 0.013177608051749928, "valid_fraction": 0.98, "identity_delta": 0.01122336469590664, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.45413086495308974, "target_raw": 2.858945999145508, "ce": 6.280273895263672, "genuineness": 4.217896122932434, "identity": 0.551513960659504, "wer": 0.022128740824392996, "valid_fraction": 0.98, "identity_delta": -0.025556373074650765, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.0038281327013279334, "ci95_z": [ -0.08579100154321143, 0.09344726694586729 ], "language_deltas": { "en": 0.004345714206852693, "de": 0.1358398728090804, "fr": -0.14379880535682993, "nl": -0.07285155560462499, "es": 0.0956054374521615 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.0011119651794433594, "delta_genuineness": 0.009162998199462891, "delta_wer": -0.008951132772643068, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.5103247144637031, "ci95_z": [ 0.38474607986790865, 0.6305005061355698 ], "language_deltas": { "en": 0.39003904657795224, "de": 0.6644531622171149, "fr": 0.5670532657825671, "nl": 0.301806627510355, "es": 0.6282714702305262 }, "languages_improved": 5, "clear_gain": true, "delta_ce": 0.010325498580932617, "delta_genuineness": -0.037781782150268554, "delta_wer": 0.006756094782410572, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Establish it immediately. Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Sustain the delivery. Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Playful speech; light mischief, lively pitch and a smile that remains audible in the words. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "language": "fr", "seed": 1702, "scale": 0.0, "cfg": 1.0, "frames": 128, "reference": true } }, { "target": "vn_S_STRY_reg", "label": "Strong Storytelling Style", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_STRY_reg", "adapter_sha256": "c7d9291ccbaba275dc8820f79e0f653b4798d05d25e26a6a3181763a93cebf2b", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_STRY_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_STRY_reg", "category": "talking_style", "head": "vn_S_STRY_reg", "pole": "high", "label": "Strong Storytelling Style", "floor": 4.0, "canonical_dimension_name": "Storytelling Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-d04a18760779", "general": "An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Establish it immediately. An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Sustain the delivery. An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.", "An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.", "An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.3034472693555547, "target_raw": 2.742946677207947, "ce": 6.231029348373413, "genuineness": 4.3371408605575565, "identity": 0.5548396833240986, "wer": 0.006331639375117636, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.761953123944151, "target_raw": 3.3819459009170534, "ce": 6.281544761657715, "genuineness": 4.307769055366516, "identity": 0.5677441400289536, "wer": 0.018921513269339354, "valid_fraction": 0.98, "identity_delta": 0.012904456704854966, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.7532812571904647, "target_raw": 3.369860305786133, "ce": 6.292664480209351, "genuineness": 4.307518005371094, "identity": 0.5779004135727882, "wer": 0.024865919740061616, "valid_fraction": 0.98, "identity_delta": 0.02306073024868965, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.761953123944151, "target_raw": 3.3819459009170534, "ce": 6.281544761657715, "genuineness": 4.307769055366516, "identity": 0.5677441400289536, "wer": 0.018921513269339354, "valid_fraction": 0.98, "identity_delta": 0.012904456704854966, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.008671866753686431, "ci95_z": [ -0.076015615657742, 0.055937523683091794 ], "language_deltas": { "en": 0.02890624532218843, "de": 0.08984378354659164, "fr": 0.035546903000043616, "nl": -0.08867190781150695, "es": -0.10898435782574889 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.011119718551635743, "delta_genuineness": -0.0002510499954223633, "delta_wer": 0.005944406470722261, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.008671866753686431, "ci95_z": [ -0.076015615657742, 0.055937523683091794 ], "language_deltas": { "en": 0.02890624532218843, "de": 0.08984378354659164, "fr": 0.035546903000043616, "nl": -0.08867190781150695, "es": -0.10898435782574889 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.011119718551635743, "delta_genuineness": -0.0002510499954223633, "delta_wer": 0.005944406470722261, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Establish it immediately. An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Sustain the delivery. An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] The road turns left behind the bridge.\n(An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(An engaging storyteller; curiosity, rhythmic suspense and an intimate awareness of the listener. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "language": "en", "seed": 1701, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_VULN_reg_high", "label": "emotionally vulnerable and exposed", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_VULN_reg_high", "adapter_sha256": "3718df2957f81e89e67607c286f7c2eb8d577ecf4233e45f944892c447e32102", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_VULN_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_VULN_reg_high", "category": "bipolar", "head": "vn_VULN_reg", "pole": "high", "label": "emotionally vulnerable and exposed", "floor": 4.0, "canonical_dimension_name": "Vulnerability" }, "configuration": { "adapter": true, "scale": 0.75, "cfg": 5.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "canonical-extension-P07", "general": "Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice.", "inline": [ "Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice.", "Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice.", "Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice." ], "recipe": "P07", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": -0.4351220675185388, "target_raw": 2.2935844254493714, "ce": 6.220862693786621, "genuineness": 4.471824707984925, "identity": 0.5991713523864746, "wer": 0.014042160737812911, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.4760546944224655, "target_raw": 4.1866492700576785, "ce": 6.004424886703491, "genuineness": 4.688097057342529, "identity": 0.6488002383708954, "wer": 0.01760926030491248, "valid_fraction": 0.98, "identity_delta": 0.04962888598442078, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.636992166610373, "target_raw": 4.346061563491821, "ce": 5.894974870681763, "genuineness": 4.6216965675354, "identity": 0.5971784099936486, "wer": 0.018901844532279315, "valid_fraction": 1, "identity_delta": -0.0019929423928260804, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.4760546944224655, "target_raw": 4.1866492700576785, "ce": 6.004424886703491, "genuineness": 4.688097057342529, "identity": 0.6488002383708954, "wer": 0.01760926030491248, "valid_fraction": 0.98, "identity_delta": 0.04962888598442078, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.16093747218790747, "ci95_z": [ 0.06781247643776066, 0.25390620091043437 ], "language_deltas": { "en": 0.11796870435168717, "de": 0.24921870999308535, "fr": 0.14609376057762138, "nl": 0.30078119673579984, "es": -0.009375010718656407 }, "languages_improved": 4, "clear_gain": true, "delta_ce": -0.10945001602172852, "delta_genuineness": -0.06640048980712891, "delta_wer": 0.0012925842273668361, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.16093747218790747, "ci95_z": [ 0.06781247643776066, 0.25390620091043437 ], "language_deltas": { "en": 0.11796870435168717, "de": 0.24921870999308535, "fr": 0.14609376057762138, "nl": 0.30078119673579984, "es": -0.009375010718656407 }, "languages_improved": 4, "clear_gain": true, "delta_ce": -0.10945001602172852, "delta_genuineness": -0.06640048980712891, "delta_wer": 0.0012925842273668361, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice.\nSCRIPT:\n(Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice.) [3.2 seconds duration] Someone moved the little blue chair.\n(Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Audibly vulnerable and emotionally exposed, with fragile breath, open feeling and tentative emphasis. React naturally to a silent partner and keep one consistent adult voice.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Someone moved the little blue chair.\n(Natural clear speech.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Natural clear speech.) [3.2 seconds duration] We can put it back after lunch.", "language": "en", "seed": 1701, "scale": 0.75, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_VULN_reg_low", "label": "emotionally detached and unexposed", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_VULN_reg_low", "adapter_sha256": "6f1243f82f9d90c07de6682960ac3b9a31568b821da8a66a1ef1d94714255fdf", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_VULN_reg_low.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_VULN_reg_low", "category": "bipolar", "head": "vn_VULN_reg", "pole": "low", "label": "emotionally detached and unexposed", "floor": 1.0, "canonical_dimension_name": "Vulnerability" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P07", "general": "Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.", "inline": [ "Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.", "Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.", "Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice." ], "recipe": "P07", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 0.6791113441083736, "target_raw": 2.0519073963165284, "ce": 6.221180419921875, "genuineness": 4.417976298332214, "identity": 0.600748625099659, "wer": 0.020072117051522086, "valid_fraction": 0.96, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.8305468702095128, "target_raw": 1.901907000541687, "ce": 6.246438093185425, "genuineness": 4.328730993270874, "identity": 0.5679080349206924, "wer": 0.008405797101449276, "valid_fraction": 1, "identity_delta": -0.032840590178966525, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.8716845708093544, "target_raw": 1.861159155368805, "ce": 6.244849576950073, "genuineness": 4.319568004608154, "identity": 0.5830864709615707, "wer": 0.006536796536796536, "valid_fraction": 1, "identity_delta": -0.017662154138088228, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.8305468702095128, "target_raw": 1.901907000541687, "ce": 6.246438093185425, "genuineness": 4.328730993270874, "identity": 0.5679080349206924, "wer": 0.008405797101449276, "valid_fraction": 1, "identity_delta": -0.032840590178966525, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.041137700599841585, "ci95_z": [ -0.0814257627215332, 0.15260743493542422 ], "language_deltas": { "en": 0.08583978995104229, "de": -0.17990719945293784, "fr": -0.015771471567263433, "nl": 0.23164063108800878, "es": 0.08388675298035814 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.0015885162353515626, "delta_genuineness": -0.009162988662719727, "delta_wer": -0.0018690005646527386, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.041137700599841585, "ci95_z": [ -0.0814257627215332, 0.15260743493542422 ], "language_deltas": { "en": 0.08583978995104229, "de": -0.17990719945293784, "fr": -0.015771471567263433, "nl": 0.23164063108800878, "es": 0.08388675298035814 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.0015885162353515626, "delta_genuineness": -0.009162988662719727, "delta_wer": -0.0018690005646527386, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.\nSCRIPT:\n(Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "language": "fr", "seed": 1701, "scale": 0.0, "cfg": 1.0, "frames": 128, "reference": true } }, { "target": "vn_STNC_reg_high", "label": "dominant and assertive", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_STNC_reg_high", "adapter_sha256": "aad1c7ea7948d09fb44a918f5cdd4d31dbead441ef62eb9712a15aca48b5df5e", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_STNC_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_STNC_reg_high", "category": "bipolar", "head": "vn_STNC_reg", "pole": "high", "label": "dominant and assertive", "floor": 4.0, "canonical_dimension_name": "Stance" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P01", "general": "Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear.", "inline": [ "Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear.", "Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear.", "Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear." ], "recipe": "P01", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 0.9617187613199555, "target_raw": 3.669060049057007, "ce": 6.240560550689697, "genuineness": 4.161286287307739, "identity": 0.5651051919162273, "wer": 0.010267268962921136, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.7081249861281471, "target_raw": 4.5189579439163206, "ce": 6.157480163574219, "genuineness": 3.9006423282623293, "identity": 0.5386939653754235, "wer": 0.01239462292093871, "valid_fraction": 1, "identity_delta": -0.026411226540803908, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.9804687660311509, "target_raw": 4.829063081741333, "ce": 6.117925672531128, "genuineness": 3.796209011077881, "identity": 0.4837169282138348, "wer": 0.005631469979296066, "valid_fraction": 1, "identity_delta": -0.08138826370239258, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.9214062394652438, "target_raw": 4.76181134223938, "ce": 6.130633926391601, "genuineness": 3.825894694328308, "identity": 0.49277658216655257, "wer": 0.0133507682249101, "valid_fraction": 1, "identity_delta": -0.0723286097496748, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.05906252656590714, "ci95_z": [ -0.006875011777988352, 0.12281256268507149 ], "language_deltas": { "en": -0.04218751897564792, "de": 0.07812510338456462, "fr": 0.167187550383617, "nl": 0.10078130463677941, "es": -0.008593806599777421 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.012708253860473633, "delta_genuineness": -0.029685683250427246, "delta_wer": -0.007719298245614035, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.2723437799030038, "ci95_z": [ 0.17484380817017592, 0.37546874718636947 ], "language_deltas": { "en": 0.1828125543096131, "de": 0.3773437149932012, "fr": 0.4390624888763443, "nl": 0.2164063504400675, "es": 0.14609379089579294 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.03955449104309082, "delta_genuineness": -0.10443331718444825, "delta_wer": -0.006763152941642644, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear.\nSCRIPT:\n(Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Dominant, assertive interpersonal stance, with decisive attacks and confident ownership of each statement. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "vn_STNC_reg_low", "label": "submissive and deferential", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_STNC_reg_low", "adapter_sha256": "0d44c29174f2c5668db6534c9a640fe939c71563135b3c7b7744387d78706bb1", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_STNC_reg_low.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_STNC_reg_low", "category": "bipolar", "head": "vn_STNC_reg", "pole": "low", "label": "submissive and deferential", "floor": 1.0, "canonical_dimension_name": "Stance" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P04", "general": "Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis.", "inline": [ "Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis.", "Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis.", "Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis." ], "recipe": "P04", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": -0.25194213060773585, "target_raw": 2.8608705615997314, "ce": 6.205930471420288, "genuineness": 4.440946626663208, "identity": 0.59952592253685, "wer": 0.006666666666666666, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.9128515525377848, "target_raw": 1.5345744156837464, "ce": 5.731196994781494, "genuineness": 4.596090307235718, "identity": 0.6139840245246887, "wer": 0.024463579898362505, "valid_fraction": 0.98, "identity_delta": 0.014458101987838746, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.9827343694054556, "target_raw": 1.4550021171569825, "ce": 5.713762836456299, "genuineness": 4.586174192428589, "identity": 0.6285103705525398, "wer": 0.02752656344418358, "valid_fraction": 0.98, "identity_delta": 0.028984448015689848, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.9128515525377848, "target_raw": 1.5345744156837464, "ce": 5.731196994781494, "genuineness": 4.596090307235718, "identity": 0.6139840245246887, "wer": 0.024463579898362505, "valid_fraction": 0.98, "identity_delta": 0.014458101987838746, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.06988281686767069, "ci95_z": [ -0.015312511385388777, 0.16484376210515472 ], "language_deltas": { "en": 0.11523437254625242, "de": 0.03671875294449709, "fr": 0.03437500654332688, "nl": -0.07617189413923113, "es": 0.23925784644350817 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.017434158325195313, "delta_genuineness": -0.009916114807128906, "delta_wer": 0.0030629835458210746, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.06988281686767069, "ci95_z": [ -0.015312511385388777, 0.16484376210515472 ], "language_deltas": { "en": 0.11523437254625242, "de": 0.03671875294449709, "fr": 0.03437500654332688, "nl": -0.07617189413923113, "es": 0.23925784644350817 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.017434158325195313, "delta_genuineness": -0.009916114807128906, "delta_wer": 0.0030629835458210746, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis.\nSCRIPT:\n(Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Meek and deferential interpersonal stance, with tentative attacks and a yielding, submissive attitude. Think while speaking, with small organic hesitations and spontaneous emphasis.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Natural clear speech.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "language": "fr", "seed": 1702, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_AROU_reg_high", "label": "highly aroused and activated", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_AROU_reg_high", "adapter_sha256": "a138d4bd5b71a5c34c2e937aba85670579ac47f5a94348700e7913c76282bedb", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_AROU_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_AROU_reg_high", "category": "bipolar", "head": "vn_AROU_reg", "pole": "high", "label": "highly aroused and activated", "floor": 4.0, "canonical_dimension_name": "Arousal" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.25, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P01", "general": "Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear.", "inline": [ "Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear.", "Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear.", "Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear." ], "recipe": "P01", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 0.03347169290207476, "target_raw": 2.4743812084198, "ce": 6.268995313644409, "genuineness": 4.313166446685791, "identity": 0.5796557331085205, "wer": 0.008362507058159233, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.5739746106223111, "target_raw": 3.083273730278015, "ce": 6.25660475730896, "genuineness": 4.226180548667908, "identity": 0.5397220122814178, "wer": 0.004679089026915113, "valid_fraction": 1, "identity_delta": -0.03993372082710266, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.874941430251698, "target_raw": 3.422321753501892, "ce": 6.186073789596557, "genuineness": 4.077626657485962, "identity": 0.5110674953460693, "wer": 0.006670995670995671, "valid_fraction": 1, "identity_delta": -0.06858823776245117, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.7607617088438484, "target_raw": 3.293694920539856, "ce": 6.235794935226441, "genuineness": 4.161600065231323, "identity": 0.5461865630745888, "wer": 0.004679089026915113, "valid_fraction": 1, "identity_delta": -0.03346917003393173, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.11417972140784963, "ci95_z": [ -0.021054667108994004, 0.23460941808028876 ], "language_deltas": { "en": 0.12578130299264076, "de": 0.15898442489244724, "fr": -0.11894531384341718, "nl": 0.32109379720561293, "es": 0.08398439579196439 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.049721145629882814, "delta_genuineness": -0.08397340774536133, "delta_wer": 0.001991906644080557, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.3009668196293868, "ci95_z": [ 0.2241992293133926, 0.37056642555593833 ], "language_deltas": { "en": 0.3709961306939724, "de": 0.3004395019478805, "fr": 0.16054687743881885, "nl": 0.40449220450972834, "es": 0.26835938355653405 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.07053096771240235, "delta_genuineness": -0.1485538911819458, "delta_wer": 0.001991906644080557, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear.\nSCRIPT:\n(Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Highly activated and alert delivery, energetic breath, lively pitch motion and urgent forward momentum. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "language": "de", "seed": 1702, "scale": 0.0, "cfg": 1.25, "frames": 128, "reference": true } }, { "target": "vn_AROU_reg_low", "label": "very low arousal, calm or sleepy", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_AROU_reg_low", "adapter_sha256": "d98bc252644859c4d336217e38d80116f4f6414d1cceda61ddf66e466b02121e", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_AROU_reg_low.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_AROU_reg_low", "category": "bipolar", "head": "vn_AROU_reg", "pole": "low", "label": "very low arousal, calm or sleepy", "floor": 1.0, "canonical_dimension_name": "Arousal" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P05", "general": "Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end.", "inline": [ "Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end.", "Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end.", "Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end." ], "recipe": "P05", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 0.5274609320725946, "target_raw": 1.8424740171432494, "ce": 6.24469072341919, "genuineness": 4.43768307685852, "identity": 0.6112351068854331, "wer": 0.013181372403340367, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.1968749933201164, "target_raw": 1.0883592796325683, "ce": 5.942710332870483, "genuineness": 4.560442419052124, "identity": 0.606009187400341, "wer": 0.007366836062488236, "valid_fraction": 1, "identity_delta": -0.005225919485092163, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.1817968691054987, "target_raw": 1.1053452324867248, "ce": 5.852322769165039, "genuineness": 4.488895568847656, "identity": 0.5997273537516594, "wer": 0.019982119329945418, "valid_fraction": 1, "identity_delta": -0.011507753133773804, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.1968749933201164, "target_raw": 1.0883592796325683, "ce": 5.942710332870483, "genuineness": 4.560442419052124, "identity": 0.606009187400341, "wer": 0.007366836062488236, "valid_fraction": 1, "identity_delta": -0.005225919485092163, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.015078124214617652, "ci95_z": [ -0.05187499490741547, 0.011328130166989147 ], "language_deltas": { "en": 0.02382813148973837, "de": 0.0007812368551795678, "fr": -0.08515624578373682, "nl": -0.012890625041335913, "es": -0.001953118592933467 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.09038756370544433, "delta_genuineness": -0.07154685020446777, "delta_wer": 0.01261528326745718, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.015078124214617652, "ci95_z": [ -0.05187499490741547, 0.011328130166989147 ], "language_deltas": { "en": 0.02382813148973837, "de": 0.0007812368551795678, "fr": -0.08515624578373682, "nl": -0.012890625041335913, "es": -0.001953118592933467 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.09038756370544433, "delta_genuineness": -0.07154685020446777, "delta_wer": 0.01261528326745718, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end.\nSCRIPT:\n(Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Very low activation, calm and settled delivery, easy breath and restrained energy with clear voiced words. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "language": "fr", "seed": 1702, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_VOLT_reg_high", "label": "Emotionally volatile delivery with rapid or abrupt changes", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_VOLT_reg_high", "adapter_sha256": "5f4f8d843177507fdb9593d7a22eb2ebf5467391d66fa5e2228a26b608413238", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_VOLT_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_VOLT_reg_high", "category": "bipolar", "head": "vn_VOLT_reg", "pole": "high", "label": "Emotionally volatile delivery with rapid or abrupt changes", "floor": 4.0, "canonical_dimension_name": "Volatility" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 3.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P01", "general": "Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear.", "inline": [ "Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear.", "Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear.", "Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear." ], "recipe": "P01", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": -1.228749995109911, "target_raw": 1.6172904920578004, "ce": 6.227852258682251, "genuineness": 4.338521647453308, "identity": 0.5755038475990295, "wer": 0.0028138528138528136, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.6421484189021123, "target_raw": 2.126728608608246, "ce": 6.166614246368408, "genuineness": 4.1675622940063475, "identity": 0.5294986745715141, "wer": 0.022006399397703744, "valid_fraction": 0.96, "identity_delta": -0.04600517302751541, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.2700488054902419, "target_raw": 2.449881041049957, "ce": 6.031350545883178, "genuineness": 3.9640930843353273, "identity": 0.5147908115386963, "wer": 0.015333333333333332, "valid_fraction": 1, "identity_delta": -0.06071303606033325, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.3799218779973243, "target_raw": 2.3544610261917116, "ce": 6.086949224472046, "genuineness": 4.0348867988586425, "identity": 0.5584957027435302, "wer": 0.011714191605495953, "valid_fraction": 1, "identity_delta": -0.017008144855499268, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.10987307250708239, "ci95_z": [ 0.028925807707150853, 0.2063769670425847 ], "language_deltas": { "en": 0.06923832042567175, "de": -0.015624963324332603, "fr": 0.053320322379373855, "nl": 0.14702146910352556, "es": 0.2954102139511734 }, "languages_improved": 4, "clear_gain": true, "delta_ce": -0.05559867858886719, "delta_genuineness": -0.07079371452331543, "delta_wer": 0.00361914172783738, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.3720996134118703, "ci95_z": [ 0.30787109958781184, 0.43632812723592884 ], "language_deltas": { "en": 0.3042968885925171, "de": 0.30429687486595147, "fr": 0.32216797102212197, "nl": 0.4727050331952967, "es": 0.45703129938346443 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.13526370048522948, "delta_genuineness": -0.2034692096710205, "delta_wer": -0.006673066064370412, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear.\nSCRIPT:\n(Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Emotionally volatile delivery with abrupt changes of attitude, pitch energy and emotional control across the connected thoughts. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "language": "de", "seed": 1701, "scale": 0.0, "cfg": 3.0, "frames": 128, "reference": true } }, { "target": "vn_VOLT_reg_low", "label": "Emotionally steady, stable delivery without abrupt changes", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_VOLT_reg_low", "adapter_sha256": "947da1ebf84d77b7f2074c287a797ff7f722ba207146fd9df1cf77d0f05bd57e", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_VOLT_reg_low.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_VOLT_reg_low", "category": "bipolar", "head": "vn_VOLT_reg", "pole": "low", "label": "Emotionally steady, stable delivery without abrupt changes", "floor": 1.0, "canonical_dimension_name": "Volatility" }, "configuration": { "adapter": true, "scale": 1.0, "cfg": 5.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "canonical-extension-P05", "general": "Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.", "inline": [ "Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.", "Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.", "Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end." ], "recipe": "P05", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 1.530625004160865, "target_raw": 1.355125093460083, "ce": 6.230711612701416, "genuineness": 4.354337244033814, "identity": 0.5973480761051178, "wer": 0.007453416149068323, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.7553124904128519, "target_raw": 1.1599937224388122, "ce": 6.205771636962891, "genuineness": 4.298606038093567, "identity": 0.6453817552328109, "wer": 0.009536796536796536, "valid_fraction": 1, "identity_delta": 0.04803367912769318, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.8610937478445002, "target_raw": 1.068127272129059, "ce": 6.205930452346802, "genuineness": 4.256305503845215, "identity": 0.6409243163466454, "wer": 0.009454664328806204, "valid_fraction": 1, "identity_delta": 0.04357624024152756, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.7553124904128519, "target_raw": 1.1599937224388122, "ce": 6.205771636962891, "genuineness": 4.298606038093567, "identity": 0.6453817552328109, "wer": 0.009536796536796536, "valid_fraction": 1, "identity_delta": 0.04803367912769318, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.10578125743164839, "ci95_z": [ 0.05921876212870756, 0.14937500956570038 ], "language_deltas": { "en": 0.14765625412869357, "de": 0.15468752262738542, "fr": 0.052343745871306455, "nl": 0.14218749431634392, "es": 0.032031270214512596 }, "languages_improved": 5, "clear_gain": true, "delta_ce": 0.00015881538391113281, "delta_genuineness": -0.04230053424835205, "delta_wer": -8.213220799033194e-05, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.10578125743164839, "ci95_z": [ 0.05921876212870756, 0.14937500956570038 ], "language_deltas": { "en": 0.14765625412869357, "de": 0.15468752262738542, "fr": 0.052343745871306455, "nl": 0.14218749431634392, "es": 0.032031270214512596 }, "languages_improved": 5, "clear_gain": true, "delta_ce": 0.00015881538391113281, "delta_genuineness": -0.04230053424835205, "delta_wer": -8.213220799033194e-05, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.\nSCRIPT:\n(Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Emotionally stable and steady delivery, maintaining a consistent state while allowing natural expressive phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "language": "es", "seed": 1702, "scale": 1.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_VALN_reg_high", "label": "strongly positive vocal affect", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_VALN_reg_high", "adapter_sha256": "641b374f50377a1a69a2f987ca5557b5397d62af0ee76c2375fe667884c5fe71", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_VALN_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_VALN_reg_high", "category": "bipolar", "head": "vn_VALN_reg", "pole": "high", "label": "strongly positive vocal affect", "floor": 4.0, "canonical_dimension_name": "Valence" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.5, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P06", "general": "Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible.", "inline": [ "Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible.", "Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible.", "Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible." ], "recipe": "P06", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 1.0292968801764066, "target_raw": 3.3137970876693728, "ce": 6.239925107955933, "genuineness": 4.356596603393554, "identity": 0.5473904445767402, "wer": 0.011195840391492564, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.6182812589021007, "target_raw": 3.979277324676514, "ce": 6.284404077529907, "genuineness": 4.157520637512207, "identity": 0.5377974936366081, "wer": 0.009232072275550536, "valid_fraction": 1, "identity_delta": -0.009592950940132142, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.727031257319505, "target_raw": 4.102151508331299, "ce": 6.258193273544311, "genuineness": 4.114341473579406, "identity": 0.5213103777170182, "wer": 0.008110295501599848, "valid_fraction": 1, "identity_delta": -0.026080066859722136, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.7472656376277946, "target_raw": 4.125013880729675, "ce": 6.266135959625244, "genuineness": 4.179737811088562, "identity": 0.5192924910783767, "wer": 0.006453980801806888, "valid_fraction": 1, "identity_delta": -0.028097953498363495, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.020234380308289637, "ci95_z": [ -0.1115625114738186, 0.09140623542693156 ], "language_deltas": { "en": -0.07109370021417795, "de": 0.03046872263428322, "fr": 0.18281250461590406, "nl": -0.07539063736402862, "es": -0.1679687912134289 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.007942686080932617, "delta_genuineness": -0.06539633750915527, "delta_wer": 0.0016563146997929607, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.10874999841740435, "ci95_z": [ 0.06156247349152255, 0.17375003850982795 ], "language_deltas": { "en": 0.05781249406526624, "de": 0.051562443949736705, "fr": 0.2328125679197308, "nl": 0.08906249142760685, "es": 0.11249999472468113 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.026210803985595704, "delta_genuineness": -0.043179163932800295, "delta_wer": -0.001121776773950687, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible.\nSCRIPT:\n(Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Strongly positive vocal affect, with bright, pleased, welcoming emotional coloring throughout. Begin with the quality contained, then let it become clearly audible.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "language": "fr", "seed": 1701, "scale": 0.0, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "vn_VALN_reg_low", "label": "strongly negative vocal affect", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_VALN_reg_low", "adapter_sha256": "eba00ec2bfac8ac72ddc3ed765f202acd8ee12cc458111b53e8f630d5009a909", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_VALN_reg_low.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_VALN_reg_low", "category": "bipolar", "head": "vn_VALN_reg", "pole": "low", "label": "strongly negative vocal affect", "floor": 1.0, "canonical_dimension_name": "Valence" }, "configuration": { "adapter": true, "scale": 1.0, "cfg": 5.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "canonical-extension-P02", "general": "Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "inline": [ "Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality." ], "recipe": "P02", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": -0.774765621801838, "target_raw": 3.0262079286575316, "ce": 6.244531888961792, "genuineness": 4.381449761390686, "identity": 0.5794402173161507, "wer": 0.005714285714285714, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.22335447821820828, "target_raw": 1.898454418182373, "ce": 6.077020902633667, "genuineness": 4.440193486213684, "identity": 0.5755717204511166, "wer": 0.01266704310182571, "valid_fraction": 1, "identity_delta": -0.0038684968650341033, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.5312402316920057, "target_raw": 1.5505812120437623, "ce": 5.977340393066406, "genuineness": 4.248272142410278, "identity": 0.6351269596815109, "wer": 0.028611989459815548, "valid_fraction": 0.96, "identity_delta": 0.05568674236536026, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.22335447821820828, "target_raw": 1.898454418182373, "ce": 6.077020902633667, "genuineness": 4.440193486213684, "identity": 0.5755717204511166, "wer": 0.01266704310182571, "valid_fraction": 1, "identity_delta": -0.0038684968650341033, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.3078857534737975, "ci95_z": [ 0.15748538420147498, 0.4460937487471117 ], "language_deltas": { "en": 0.07136235110412426, "de": 0.4061523512096306, "fr": 0.5234374927464365, "nl": 0.1671875164853715, "es": 0.37128905582342453 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.09968050956726074, "delta_genuineness": -0.19192134380340575, "delta_wer": 0.015944946357989837, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.3078857534737975, "ci95_z": [ 0.15748538420147498, 0.4460937487471117 ], "language_deltas": { "en": 0.07136235110412426, "de": 0.4061523512096306, "fr": 0.5234374927464365, "nl": 0.1671875164853715, "es": 0.37128905582342453 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.09968050956726074, "delta_genuineness": -0.19192134380340575, "delta_wer": 0.015944946357989837, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality.\nSCRIPT:\n(Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Strongly negative vocal affect, with grim, displeased, disheartened emotional coloring throughout. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "language": "nl", "seed": 1701, "scale": 1.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_WARM_reg_high", "label": "very warm vocal delivery", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_WARM_reg_high", "adapter_sha256": "df5732bc9c39127cac838ba98c6f197ad8648d254b4082d2b504bf0767e0025c", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_WARM_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_WARM_reg_high", "category": "bipolar", "head": "vn_WARM_reg", "pole": "high", "label": "very warm vocal delivery", "floor": 4.0, "canonical_dimension_name": "Warmth" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P05", "general": "Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end.", "inline": [ "Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end.", "Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end.", "Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end." ], "recipe": "P05", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 1.0912500092220878, "target_raw": 3.592702703475952, "ce": 6.265659379959106, "genuineness": 4.3132919454574585, "identity": 0.5905555510520935, "wer": 0.006680337206652996, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.4284375253607418, "target_raw": 3.897808437347412, "ce": 6.231664724349976, "genuineness": 4.322078394889831, "identity": 0.6065105706453323, "wer": 0.02362417902463669, "valid_fraction": 0.96, "identity_delta": 0.01595501959323883, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.4265624588299648, "target_raw": 3.89611177444458, "ce": 6.2015620136260985, "genuineness": 4.27011281490326, "identity": 0.5929865968227387, "wer": 0.02902552824749621, "valid_fraction": 0.94, "identity_delta": 0.0024310457706451416, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.4284375253607418, "target_raw": 3.897808437347412, "ce": 6.231664724349976, "genuineness": 4.322078394889831, "identity": 0.6065105706453323, "wer": 0.02362417902463669, "valid_fraction": 0.96, "identity_delta": 0.01595501959323883, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.0018750665307769884, "ci95_z": [ -0.05828137696838877, 0.05234372776818093 ], "language_deltas": { "en": 0.037499907779121024, "de": 0.0890625247020211, "fr": -0.09531264985892839, "nl": -0.04296884304427967, "es": 0.0023437277681809922 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.030102710723876953, "delta_genuineness": -0.051965579986572266, "delta_wer": 0.00540134922285952, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.0018750665307769884, "ci95_z": [ -0.05828137696838877, 0.05234372776818093 ], "language_deltas": { "en": 0.037499907779121024, "de": 0.0890625247020211, "fr": -0.09531264985892839, "nl": -0.04296884304427967, "es": 0.0023437277681809922 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.030102710723876953, "delta_genuineness": -0.051965579986572266, "delta_wer": 0.00540134922285952, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end.\nSCRIPT:\n(Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Very warm and personally engaged delivery, with kind resonance and a sympathetic interpersonal tone. Establish the delivery, sustain it through the middle thought and release it at the end.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "language": "fr", "seed": 1702, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_TENS_reg_high", "label": "very tense and strained", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_TENS_reg_high", "adapter_sha256": "ccfd3844e64a8833d9a4bdaaadd4a2b33b4b54459618c13ea5e42728558663f0", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_TENS_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_TENS_reg_high", "category": "bipolar", "head": "vn_TENS_reg", "pole": "high", "label": "very tense and strained", "floor": 4.0, "canonical_dimension_name": "Tension" }, "configuration": { "adapter": true, "scale": 0.75, "cfg": 3.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "canonical-extension-P02", "general": "Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "inline": [ "Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality." ], "recipe": "P02", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 0.9297656346681823, "target_raw": 2.4332660746574404, "ce": 6.235477237701416, "genuineness": 4.305760746002197, "identity": 0.5616777104139328, "wer": 0.0037267080745341614, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.5208593882489905, "target_raw": 2.92522111415863, "ce": 6.147154722213745, "genuineness": 4.31981906414032, "identity": 0.5567243963479995, "wer": 0.021119968894568436, "valid_fraction": 0.96, "identity_delta": -0.004953314065933228, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.6910937736333345, "target_raw": 3.0669036531448364, "ce": 6.020052108764649, "genuineness": 4.215134692192078, "identity": 0.579901976287365, "wer": 0.011609260304912479, "valid_fraction": 1, "identity_delta": 0.01822426587343216, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.5208593882489905, "target_raw": 2.92522111415863, "ce": 6.147154722213745, "genuineness": 4.31981906414032, "identity": 0.5567243963479995, "wer": 0.021119968894568436, "valid_fraction": 0.96, "identity_delta": -0.004953314065933228, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.1702343853843439, "ci95_z": [ 0.07249998854141362, 0.2731250429696989 ], "language_deltas": { "en": 0.07734373925757528, "de": 0.2828125214848495, "fr": 0.051953116047979385, "nl": 0.10390623209595881, "es": 0.3351563180353566 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.12710261344909668, "delta_genuineness": -0.10468437194824219, "delta_wer": -0.009510708589655957, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.1702343853843439, "ci95_z": [ 0.07249998854141362, 0.2731250429696989 ], "language_deltas": { "en": 0.07734373925757528, "de": 0.2828125214848495, "fr": 0.051953116047979385, "nl": 0.10390623209595881, "es": 0.3351563180353566 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.12710261344909668, "delta_genuineness": -0.10468437194824219, "delta_wer": -0.009510708589655957, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.\nSCRIPT:\n(Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Audibly taut and strained delivery, with pressed breath, held muscular tension and tightly controlled articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Natural clear speech.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Prends-les quand tu partiras.", "language": "fr", "seed": 1702, "scale": 0.75, "cfg": 3.0, "frames": 128, "reference": true } }, { "target": "vn_TEMP_reg_high", "label": "very fast speaking tempo", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_TEMP_reg_high", "adapter_sha256": "5b36cf57c6bcb3ca2e198656fb52cb3159df02a8bcf4f871e889ddca267a51f8", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_TEMP_reg_high.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_TEMP_reg_high", "category": "bipolar", "head": "vn_TEMP_reg", "pole": "high", "label": "very fast speaking tempo", "floor": 4.0, "canonical_dimension_name": "Tempo" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "canonical-extension-P01", "general": "Very fast speaking tempo, with rapid connected syllables and crisp economical articulation. Make this audible from the opening words and keep every word clear.", "inline": [ "Very fast speaking tempo, with rapid connected syllables and crisp economical articulation. Make this audible from the opening words and keep every word clear.", "Very fast speaking tempo, with rapid connected syllables and crisp economical articulation. Make this audible from the opening words and keep every word clear.", "Very fast speaking tempo, with rapid connected syllables and crisp economical articulation. Make this audible from the opening words and keep every word clear." ], "recipe": "P01", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 0.4628710974228222, "target_raw": 2.6707136058807373, "ce": 6.245326137542724, "genuineness": 4.277393040657043, "identity": 0.5485528255999088, "wer": 0.004596273291925465, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.6522656348821023, "target_raw": 2.8981121397018432, "ce": 6.106726503372192, "genuineness": 3.98078736782074, "identity": 0.48603457659482957, "wer": 0.0019047619047619048, "valid_fraction": 1, "identity_delta": -0.06251824900507927, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.7179687585324275, "target_raw": 2.976999282836914, "ce": 6.0805951023101805, "genuineness": 3.92593475818634, "identity": 0.5053402051329613, "wer": 0.006666666666666666, "valid_fraction": 0.98, "identity_delta": -0.04321262046694756, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.7144140621354327, "target_raw": 2.9727312994003294, "ce": 6.044535360336304, "genuineness": 3.918215231895447, "identity": 0.5141749057173729, "wer": 0.008541809068124857, "valid_fraction": 1, "identity_delta": -0.03437791988253593, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.0035546963969948038, "ci95_z": [ -0.01863279013728334, 0.028984365366113745 ], "language_deltas": { "en": 0.010156286767005557, "de": -0.0037109096920432847, "fr": -0.03320310925379299, "nl": -0.00625000124108036, "es": 0.050781215404885095 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.03605974197387695, "delta_genuineness": 0.007719526290893555, "delta_wer": -0.0018751424014581911, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.06570312365032512, "ci95_z": [ 0.013906255739996634, 0.11289062309183898 ], "language_deltas": { "en": 0.08769529392257844, "de": 0.1523437653583694, "fr": -0.011132788647986929, "nl": 0.08437499689729913, "es": 0.01523435072136552 }, "languages_improved": 4, "clear_gain": true, "delta_ce": -0.026131401062011717, "delta_genuineness": -0.05485260963439941, "delta_wer": 0.0047619047619047615, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "casual_not_formal", "label": "Very informal/casual, not formal", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/casual_not_formal", "adapter_sha256": "094c26ef18d6c6285cac07cde7d10b0f5e852aa6faf0941ca6453cc4564556b0", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/casual_not_formal.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "casual_not_formal", "category": "contrast", "head": "vn_S_CASU_reg", "pole": "high", "label": "Very informal/casual, not formal", "floor": 4.0, "opposite": "vn_S_FORM_reg", "canonical_dimension_name": "Casual Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 5.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P09", "general": "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.", "inline": [ "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.", "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.", "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence." ], "recipe": "P09", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": -0.9862597692964027, "target_raw": 2.703663535118103, "ce": 6.233412141799927, "genuineness": 4.391993470191956, "identity": 0.6015608149766922, "wer": 0.0076364825106243865, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.3753100617203283, "target_raw": 3.495331163406372, "ce": 6.146042718887329, "genuineness": 4.373039875030518, "identity": 0.5753106416016817, "wer": 0.022835309617918312, "valid_fraction": 1, "identity_delta": -0.02625017337501049, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.5323107915282019, "target_raw": 3.3067342925071714, "ce": 6.178607683181763, "genuineness": 4.368646602630616, "identity": 0.5903049494326115, "wer": 0.01594635798983625, "valid_fraction": 0.98, "identity_delta": -0.011255865544080734, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.4420214990178078, "target_raw": 3.399321026802063, "ce": 6.125868330001831, "genuineness": 4.390863819122314, "identity": 0.5562701150774956, "wer": 0.03873708480687885, "valid_fraction": 0.96, "identity_delta": -0.045290699899196624, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.09028929251039414, "ci95_z": [ -0.193347154719618, 0.055709246331492254 ], "language_deltas": { "en": 0.18636477493289066, "de": -0.2122619588972646, "fr": -0.1723144328837407, "nl": -0.18494871145991, "es": -0.06828613424394608 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.05273935317993164, "delta_genuineness": -0.022217216491699218, "delta_wer": -0.022790726817042608, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.15700072980787363, "ci95_z": [ -0.2092333811068628, -0.0916577055208151 ], "language_deltas": { "en": -0.18034667024236997, "de": -0.1392151321112478, "fr": -0.23530271087668778, "nl": -0.03253172903977852, "es": -0.19760740676928415 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.03256496429443359, "delta_genuineness": -0.004393272399902344, "delta_wer": -0.006888951628082063, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "language": "fr", "seed": 1701, "scale": 0.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "formal_not_casual", "label": "Very formal, not casual", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/formal_not_casual", "adapter_sha256": "7489a4ff2a6d51178e56f22077b1d5902c888187124567dfd2ecc528ed01686b", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/formal_not_casual.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "formal_not_casual", "category": "contrast", "head": "vn_S_FORM_reg", "pole": "high", "label": "Very formal, not casual", "floor": 4.0, "opposite": "vn_S_CASU_reg", "canonical_dimension_name": "Formal Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "canonical-extension-P05", "general": "Very formal, composed delivery with precise diction, measured authority and carefully structured phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.", "inline": [ "Very formal, composed delivery with precise diction, measured authority and carefully structured phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.", "Very formal, composed delivery with precise diction, measured authority and carefully structured phrasing. Establish the delivery, sustain it through the middle thought and release it at the end.", "Very formal, composed delivery with precise diction, measured authority and carefully structured phrasing. Establish the delivery, sustain it through the middle thought and release it at the end." ], "recipe": "P05", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 1.2890429717288774, "target_raw": 2.7567947459220887, "ce": 6.251997995376587, "genuineness": 4.271619086265564, "identity": 0.5999431225657463, "wer": 0.0038095238095238095, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 2.0893749976847302, "target_raw": 3.457391200065613, "ce": 6.228805389404297, "genuineness": 4.046246461868286, "identity": 0.5114372865110636, "wer": 0.013264188138330014, "valid_fraction": 0.98, "identity_delta": -0.08850583605468273, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.13710938276135, "target_raw": 3.501821002960205, "ce": 6.220227308273316, "genuineness": 4.026476902961731, "identity": 0.5223555308580399, "wer": 0.011176359872012046, "valid_fraction": 1, "identity_delta": -0.07758759170770645, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 2.0893749976847302, "target_raw": 3.457391200065613, "ce": 6.228805389404297, "genuineness": 4.046246461868286, "identity": 0.5114372865110636, "wer": 0.013264188138330014, "valid_fraction": 0.98, "identity_delta": -0.08850583605468273, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.04773438507661958, "ci95_z": [ -0.021640639953638956, 0.1242969184540964 ], "language_deltas": { "en": -0.055078157801116225, "de": 0.02773433892762418, "fr": 0.17773442400694606, "nl": -0.012890611546793251, "es": 0.10117193179643716 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.008578081130981446, "delta_genuineness": -0.019769558906555174, "delta_wer": -0.002087828266317969, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.04773438507661958, "ci95_z": [ -0.021640639953638956, 0.1242969184540964 ], "language_deltas": { "en": -0.055078157801116225, "de": 0.02773433892762418, "fr": 0.17773442400694606, "nl": -0.012890611546793251, "es": 0.10117193179643716 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.008578081130981446, "delta_genuineness": -0.019769558906555174, "delta_wer": -0.002087828266317969, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Concentration", "label": "Intense Concentration", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Concentration", "adapter_sha256": "03685ad489d5469b3266b3f8290d6fa8bd0ffeab426c7f1e3047c7dd09ddb12d", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Concentration.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Concentration", "category": "emotion", "head": "emo_Concentration", "pole": "high", "label": "Intense Concentration", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-04d8f738e9e1", "general": "Employ a highly rhythmic and measured delivery, where each syllable is placed with the precision of a complex calculation.", "inline": [ "Begin with a very low, grounded vocal placement, establishing a sense of unshakeable mental foundation.", "Maintain near-perfect, deliberate pacing throughout, using slight pauses between conceptual units for emphasis.", "Conclude with a brief, held vocal sustain on the final vowel, allowing the auditory impression of prolonged thought to linger." ], "cfg": 2.5, "recipe": "P07", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 1.8126562478711392, "target_raw": 1.994885196685791, "ce": 6.243419885635376, "genuineness": 4.395508089065552, "identity": 0.6163848227262497, "wer": 0.011176359872012046, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.9576562629643792, "target_raw": 2.100819692611694, "ce": 5.963281888961792, "genuineness": 4.42400128364563, "identity": 0.5273979195952415, "wer": 0.02538678712591756, "valid_fraction": 0.98, "identity_delta": -0.08898690313100815, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.87328126301537, "target_raw": 2.0391767811775208, "ce": 6.208472156524659, "genuineness": 4.39488046169281, "identity": 0.5744029092788696, "wer": 0.008488612836438923, "valid_fraction": 1, "identity_delta": -0.041981913447380066, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.879609364107117, "target_raw": 2.0437999820709227, "ce": 6.180831642150879, "genuineness": 4.390989351272583, "identity": 0.6089745083451271, "wer": 0.006583850931677018, "valid_fraction": 1, "identity_delta": -0.0074103143811225895, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.006328101091747082, "ci95_z": [ -0.08468746754443514, 0.08234377619645952 ], "language_deltas": { "en": 0.13437500076486017, "de": -0.06640614374818013, "fr": 0.07500002243589718, "nl": -0.1160156653145028, "es": -0.058593719596809826 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.027640514373779297, "delta_genuineness": 0.0038911104202270508, "delta_wer": 0.0019047619047619048, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.08437499994900931, "ci95_z": [ -0.15437498934294885, -0.01101563064084341 ], "language_deltas": { "en": 0.03828122099905344, "de": -0.14179682818418612, "fr": -0.028124988017191255, "nl": -0.1980468858036494, "es": -0.09218751873907323 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.24519026756286622, "delta_genuineness": -0.029120821952819825, "delta_wer": -0.016898174289478637, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Doubt", "label": "Intense Doubt", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Doubt", "adapter_sha256": "47cd83f1b2bab04d2a7396c5f036da256442de09c4f5c13f0fbb114da3ec148d", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Doubt.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Doubt", "category": "emotion", "head": "emo_Doubt", "pole": "high", "label": "Intense Doubt", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-93bbe240eddf", "general": "Focus on vocal texture to create a sense of fragile hesitation, as if the voice itself is wavering.", "inline": [ "Maintain a slightly breathy quality throughout the first phrase, implying lack of air support.", "Introduce a momentary vocal fry or slight catch in the second segment for emphasis.", "Deliver the final phrase with a thin, slightly brittle resonance." ], "cfg": 3, "recipe": "P07", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.10476562709062785, "target_raw": 0.2631469559669495, "ce": 6.261370334625244, "genuineness": 4.357475299835205, "identity": 0.596659539937973, "wer": 0.01148051948051948, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.0520800866235949, "target_raw": 0.8222972747683525, "ce": 5.91602294921875, "genuineness": 4.540735635757446, "identity": 0.5406948231905698, "wer": 0.03220929794842838, "valid_fraction": 0.94, "identity_delta": -0.055964716747403144, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.1913281265125741, "target_raw": 0.8896014839410782, "ce": 5.901646718978882, "genuineness": 4.538099708557129, "identity": 0.5350658848509192, "wer": 0.027258610954263126, "valid_fraction": 0.94, "identity_delta": -0.06159365508705378, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.0520800866235949, "target_raw": 0.8222972747683525, "ce": 5.91602294921875, "genuineness": 4.540735635757446, "identity": 0.5406948231905698, "wer": 0.03220929794842838, "valid_fraction": 0.94, "identity_delta": -0.055964716747403144, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.13924803988897916, "ci95_z": [ -0.10713867753872283, 0.3856347573166811 ], "language_deltas": { "en": 0.03500974891563276, "de": 0.4831054301046959, "fr": 0.4634765887291905, "nl": -0.17421870250324734, "es": -0.11113286580137612 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.014376230239868164, "delta_genuineness": -0.0026359272003173826, "delta_wer": -0.004950686994165255, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.13924803988897916, "ci95_z": [ -0.10713867753872283, 0.3856347573166811 ], "language_deltas": { "en": 0.03500974891563276, "de": 0.4831054301046959, "fr": 0.4634765887291905, "nl": -0.17421870250324734, "es": -0.11113286580137612 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.014376230239868164, "delta_genuineness": -0.0026359272003173826, "delta_wer": -0.004950686994165255, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Embarrassment", "label": "Intense Embarrassment", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Embarrassment", "adapter_sha256": "71de62032b2bdfe334acf23a6ac46e525f37a38dbc3eb25bf4773f873e54e3e9", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Embarrassment.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Embarrassment", "category": "emotion", "head": "emo_Embarrassment", "pole": "high", "label": "Intense Embarrassment", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-21ce2b7b89e4", "general": "Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Establish it from the first phrase. Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Sustain it through the thought. Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.", "Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.", "Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.5461328121830249, "target_raw": 0.01934436082839966, "ce": 6.277255687713623, "genuineness": 4.344797658920288, "identity": 0.5721578750014306, "wer": 0.006623376623376623, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.4518408209481149, "target_raw": 0.060554761290550235, "ce": 6.050810089111328, "genuineness": 4.673411159515381, "identity": 0.6182437577843666, "wer": 0.019592627814595778, "valid_fraction": 0.98, "identity_delta": 0.0460858827829361, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.49423828306794515, "target_raw": 0.04202491194009781, "ce": 6.097274713516235, "genuineness": 4.601613254547119, "identity": 0.6336197078227996, "wer": 0.02047677989440003, "valid_fraction": 1, "identity_delta": 0.06146183282136917, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.4661621077947862, "target_raw": 0.05429562985897064, "ce": 6.085440130233764, "genuineness": 4.6170523071289065, "identity": 0.6368197852373123, "wer": 0.03312817617165443, "valid_fraction": 0.96, "identity_delta": 0.0646619102358818, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.028076175273158973, "ci95_z": [ -0.07065430162096883, 0.003574216015757249 ], "language_deltas": { "en": -0.00976562555936773, "de": -0.01308594718238908, "fr": -0.10903320458002201, "nl": -0.02822266219661172, "es": 0.01972656315259568 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.011834583282470702, "delta_genuineness": -0.01543905258178711, "delta_wer": -0.0126513962772544, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.04239746211983025, "ci95_z": [ -0.08512207129971745, -0.002109375420857622 ], "language_deltas": { "en": -0.09003906581625155, "de": -0.01621093631467314, "fr": -0.10856933807341905, "nl": 0.012695306067271084, "es": -0.009863276462078569 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.04646462440490723, "delta_genuineness": -0.07179790496826172, "delta_wer": 0.0008841520798042528, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Establish it from the first phrase. Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Sustain it through the thought. Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.\nSCRIPT:\n(Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Escribí el número en el sobre.\n(Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "language": "es", "seed": 1701, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Emotional_Numbness", "label": "Intense Emotional Numbness", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Emotional_Numbness", "adapter_sha256": "8014fedb694c5c04128f522f2e477a2313d07d842b2b663851d012b4b194335f", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Emotional_Numbness.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Emotional_Numbness", "category": "emotion", "head": "emo_Emotional_Numbness", "pole": "high", "label": "Intense Emotional Numbness", "floor": 3.0 }, "configuration": { "adapter": true, "scale": 0.5, "cfg": 5.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "manual-53d14c32a3a7", "general": "Maintain a consistently flat, almost monotone pitch contour throughout the entire utterance, devoid of any natural inflections.", "inline": [ "Initiate with a breathy, minimal vocal projection that suggests detachment.", "Sustain a mid-range, unchanging vocal placement across the core of the statement.", "Deliver the closing words with the same low energy, ensuring no rise or fall." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.33365722441754947, "target_raw": 0.14102415651082992, "ce": 6.246914663314819, "genuineness": 4.376679911613464, "identity": 0.5939327973127365, "wer": 0.01784490871447393, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.11224365009981571, "target_raw": 0.5697427728772163, "ce": 5.808558654785156, "genuineness": 4.639771585464477, "identity": 0.6012379646301269, "wer": 0.025617824204780724, "valid_fraction": 0.96, "identity_delta": 0.007305167317390442, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.15911132580145493, "target_raw": 0.614804458618164, "ce": 5.848351392745972, "genuineness": 4.647177286148072, "identity": 0.6096303921937942, "wer": 0.022554607863532347, "valid_fraction": 1, "identity_delta": 0.01569759488105774, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.11224365009981571, "target_raw": 0.5697427728772163, "ce": 5.808558654785156, "genuineness": 4.639771585464477, "identity": 0.6012379646301269, "wer": 0.025617824204780724, "valid_fraction": 0.96, "identity_delta": 0.007305167317390442, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.04686767570163923, "ci95_z": [ 0.013994134857910307, 0.10048585041987572 ], "language_deltas": { "en": 0.02226561904281308, "de": 0.003686514171773797, "fr": 0.03190917600360952, "nl": 0.14947511351937912, "es": 0.027001955770620624 }, "languages_improved": 5, "clear_gain": true, "delta_ce": 0.03979273796081543, "delta_genuineness": 0.00740570068359375, "delta_wer": -0.003063216341248378, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.04686767570163923, "ci95_z": [ 0.013994134857910307, 0.10048585041987572 ], "language_deltas": { "en": 0.02226561904281308, "de": 0.003686514171773797, "fr": 0.03190917600360952, "nl": 0.14947511351937912, "es": 0.027001955770620624 }, "languages_improved": 5, "clear_gain": true, "delta_ce": 0.03979273796081543, "delta_genuineness": 0.00740570068359375, "delta_wer": -0.003063216341248378, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Maintain a consistently flat, almost monotone pitch contour throughout the entire utterance, devoid of any natural inflections.\nSCRIPT:\n(Initiate with a breathy, minimal vocal projection that suggests detachment.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Sustain a mid-range, unchanging vocal placement across the core of the statement.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Deliver the closing words with the same low energy, ensuring no rise or fall.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "language": "de", "seed": 1702, "scale": 0.5, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "emo_Infatuation", "label": "Intense Infatuation", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Infatuation", "adapter_sha256": "0832e62b6cf1a770eebdd89c7cacc24572b5dc2c540286837ad4191566fc148e", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Infatuation.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Infatuation", "category": "emotion", "head": "emo_Infatuation", "pole": "high", "label": "Intense Infatuation", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.5, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-e0a2c5b12b00", "general": "Manipulate vocal projection subtly; the emotion should be felt more than loudly stated, a warm intimacy.", "inline": [ "Begin at a hushed, intimate volume, as if sharing a secret only with the listener.", "Allow the volume to swell ever so slightly mid-phrase, a momentary rush of feeling.", "Bring the volume back down to a near whisper for the final words, sealing the private moment." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.27502929767621803, "target_raw": 0.04958263874053955, "ce": 6.275190582275391, "genuineness": 4.376303386688233, "identity": 0.6062064543366432, "wer": 0.010227743271221532, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.01589599393772067, "target_raw": 0.1898345983028412, "ce": 6.059547004699707, "genuineness": 4.504334597587586, "identity": 0.5745876064896583, "wer": 0.008196875588179936, "valid_fraction": 1, "identity_delta": -0.03161884784698486, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.04286377144960257, "target_raw": 0.17523869886994362, "ce": 5.868843479156494, "genuineness": 4.4070560121536255, "identity": 0.5307777586579323, "wer": 0.012811148424420736, "valid_fraction": 1, "identity_delta": -0.07542869567871094, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.030187994477784218, "target_raw": 0.21477685615420342, "ce": 5.879010124206543, "genuineness": 4.427013759613037, "identity": 0.5374657508730888, "wer": 0.03558876935421558, "valid_fraction": 0.98, "identity_delta": -0.06874070346355438, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.0730517659273868, "ci95_z": [ -0.13878663475989475, -0.007316897094878844 ], "language_deltas": { "en": -0.09976806534019984, "de": -0.11363525122700917, "fr": 0.02186278724326745, "nl": 0.009729002689635363, "es": -0.18344730300262777 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.010166645050048828, "delta_genuineness": -0.01995774745941162, "delta_wer": -0.02277762092979484, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.0269677775118819, "ci95_z": [ -0.09407225970889208, 0.045590821144368646 ], "language_deltas": { "en": -0.14289549935669224, "de": -0.03785399546254119, "fr": 0.10122069888230853, "nl": 0.002111820244299448, "es": -0.057421911866784046 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.1907035255432129, "delta_genuineness": -0.09727858543395997, "delta_wer": 0.0046142728362408, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Manipulate vocal projection subtly; the emotion should be felt more than loudly stated, a warm intimacy.\nSCRIPT:\n(Begin at a hushed, intimate volume, as if sharing a secret only with the listener.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Allow the volume to swell ever so slightly mid-phrase, a momentary rush of feeling.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Bring the volume back down to a near whisper for the final words, sealing the private moment.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "language": "es", "seed": 1701, "scale": 0.0, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "emo_Intoxication_Altered_States_of_Consciousness", "label": "Intense Intoxication Altered States of Consciousness", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Intoxication_Altered_States_of_Consciousness", "adapter_sha256": "a1a1d1031bc04ef0c298b7ec46f8c6df2ef855530b7d583516e91276b41a964a", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Intoxication_Altered_States_of_Consciousness.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Intoxication_Altered_States_of_Consciousness", "category": "emotion", "head": "emo_Intoxication_Altered_States_of_Consciousness", "pole": "high", "label": "Intense Intoxication Altered States of Consciousness", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-85e4ed8b7889", "general": "Employ a fluctuating vocal weight, shifting from airy lightness to sudden, momentary density.", "inline": [ "Initiate with a breathy quality that seems too voluminous for the intended words.", "Introduce brief moments of vocal compression, like a sudden, unfocused thought catching.", "Let the final articulation feel slightly weighted down, as if the effort to speak is considerable." ], "cfg": 2.5, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.2781445315806924, "target_raw": 0.017755993008613587, "ce": 6.242625637054443, "genuineness": 4.375926818847656, "identity": 0.5795974230766296, "wer": 0.011393939393939394, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.21186523427662948, "target_raw": 0.054621559232473374, "ce": 6.081230525970459, "genuineness": 4.613537731170655, "identity": 0.5934282898902893, "wer": 0.02131649281306032, "valid_fraction": 0.98, "identity_delta": 0.013830866813659668, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.19067382754524181, "target_raw": 0.06640854611992836, "ce": 6.006331100463867, "genuineness": 4.593956460952759, "identity": 0.5864316427707672, "wer": 0.03685500312044934, "valid_fraction": 0.96, "identity_delta": 0.006834219694137573, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.20488281239116451, "target_raw": 0.05850528985261917, "ce": 5.9865538597106935, "genuineness": 4.556425752639771, "identity": 0.5735473555326461, "wer": 0.026866741953698474, "valid_fraction": 1, "identity_delta": -0.006050067543983459, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.014208984845922723, "ci95_z": [ -0.023652344909516402, 0.06751953573318434 ], "language_deltas": { "en": -0.028906251841831113, "de": 0.009374999330243241, "fr": 0.0035156202279830683, "nl": 0.11904297705916984, "es": -0.03198242054595143 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.019777240753173827, "delta_genuineness": 0.037530708312988284, "delta_wer": 0.009988261166750868, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.021191406731387676, "ci95_z": [ -0.003496094407198824, 0.051074220729968436 ], "language_deltas": { "en": -0.0012695319825464595, "de": 0.008398439048812517, "fr": 0.032226561620944255, "nl": 0.07827148852958493, "es": -0.01166992355985686 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.0748994255065918, "delta_genuineness": -0.01958127021789551, "delta_wer": 0.015538510307389026, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Employ a fluctuating vocal weight, shifting from airy lightness to sudden, momentary density.\nSCRIPT:\n(Initiate with a breathy quality that seems too voluminous for the intended words.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Introduce brief moments of vocal compression, like a sudden, unfocused thought catching.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Let the final articulation feel slightly weighted down, as if the effort to speak is considerable.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Jealousy_and_Envy", "label": "Intense Jealousy and Envy", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Jealousy_and_Envy", "adapter_sha256": "d01a2ddafdc3dd5fb8d80d567e520fd7515e8146d5fdc8e1c30a8673162f6b67", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Jealousy_and_Envy.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Jealousy_and_Envy", "category": "emotion", "head": "emo_Jealousy_and_Envy", "pole": "high", "label": "Intense Jealousy and Envy", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-c1306bd27b08", "general": "Adopt a heavy, resonant vocal quality that sounds deliberately slow and weighty, using pauses as weapons to emphasize underlying suspicion and control.", "inline": [ "Begin with a deep, deliberate breath that forces a slow, measured enunciation of the first phrase.", "Insert a significant, pregnant pause mid-thought, letting the silence itself communicate the weight of the perceived slight.", "Conclude the statement with a low, drawn-out vowel sound that trails off, suggesting unresolved resentment." ], "cfg": 3, "recipe": "P02", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.48046875212063095, "target_raw": 0.0004715690016746521, "ce": 6.263117752075195, "genuineness": 4.361742987632751, "identity": 0.5780346509814263, "wer": 0.005631469979296066, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.45957031462324255, "target_raw": 0.009787256121635437, "ce": 5.977022657394409, "genuineness": 4.542367429733276, "identity": 0.588259008526802, "wer": 0.02363081121776774, "valid_fraction": 0.98, "identity_delta": 0.010224357545375824, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.4619531232502183, "target_raw": 0.008725095391273498, "ce": 5.955736312866211, "genuineness": 4.519522638320923, "identity": 0.5942593382298946, "wer": 0.04115859807621821, "valid_fraction": 0.92, "identity_delta": 0.0162246872484684, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.45957031462324255, "target_raw": 0.009787256121635437, "ce": 5.977022657394409, "genuineness": 4.542367429733276, "identity": 0.588259008526802, "wer": 0.02363081121776774, "valid_fraction": 0.98, "identity_delta": 0.010224357545375824, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.0023828086269757406, "ci95_z": [ -0.008828122707002494, 0.005742194256246636 ], "language_deltas": { "en": -0.011523429182007916, "de": 0.012695324526484631, "fr": -0.00019531031930192233, "nl": -0.004687501149110357, "es": -0.00820312701094314 }, "languages_improved": 1, "clear_gain": false, "delta_ce": -0.02128634452819824, "delta_genuineness": -0.022844791412353516, "delta_wer": 0.017527786858450474, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.0023828086269757406, "ci95_z": [ -0.008828122707002494, 0.005742194256246636 ], "language_deltas": { "en": -0.011523429182007916, "de": 0.012695324526484631, "fr": -0.00019531031930192233, "nl": -0.004687501149110357, "es": -0.00820312701094314 }, "languages_improved": 1, "clear_gain": false, "delta_ce": -0.02128634452819824, "delta_genuineness": -0.022844791412353516, "delta_wer": 0.017527786858450474, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a heavy, resonant vocal quality that sounds deliberately slow and weighty, using pauses as weapons to emphasize underlying suspicion and control.\nSCRIPT:\n(Begin with a deep, deliberate breath that forces a slow, measured enunciation of the first phrase.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Insert a significant, pregnant pause mid-thought, letting the silence itself communicate the weight of the perceived slight.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Conclude the statement with a low, drawn-out vowel sound that trails off, suggesting unresolved resentment.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "language": "fr", "seed": 1701, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Malevolence_Malice", "label": "Intense Malevolence Malice", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Malevolence_Malice", "adapter_sha256": "f53cbf0e04c6b946dd25fd607069b51eeb31ce758cb002fb7bb5f3eab052c42c", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Malevolence_Malice.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Malevolence_Malice", "category": "emotion", "head": "emo_Malevolence_Malice", "pole": "high", "label": "Intense Malevolence Malice", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-11c936b4d8e6", "general": "Focus on precise articulation, making each consonant sharp and biting, suggesting intellectual cruelty over raw aggression.", "inline": [ "Deliver the opening with unnaturally crisp diction, each syllable sounding carved from ice.", "Introduce subtle vocal tension across the mid-range frequencies, suggesting restrained, powerful enjoyment.", "End with a perfectly controlled, almost surgical delivery, leaving no room for emotional bleed." ], "cfg": 2.5, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.49449218430332076, "target_raw": 0.021068395376205445, "ce": 6.298542051315308, "genuineness": 4.322454962730408, "identity": 0.5935348200798035, "wer": 0.00831545266327875, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.3413183519912934, "target_raw": 0.10128320306539536, "ce": 6.209345846176148, "genuineness": 4.218021655082703, "identity": 0.5181017422676086, "wer": 0.008536231884057971, "valid_fraction": 1, "identity_delta": -0.07543307781219483, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.4451220672997743, "target_raw": 0.046922775208950045, "ce": 6.253427677154541, "genuineness": 4.219778957366944, "identity": 0.5478835234045982, "wer": 0.012042160737812911, "valid_fraction": 1, "identity_delta": -0.04565129667520523, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.3825976525309149, "target_raw": 0.07966586083173752, "ce": 6.244373044967651, "genuineness": 4.213000855445862, "identity": 0.5576047718524932, "wer": 0.00658008658008658, "valid_fraction": 1, "identity_delta": -0.035930048227310184, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.06252441476885941, "ci95_z": [ -0.18345703828136056, 0.012543947608307024 ], "language_deltas": { "en": -0.009375003201125259, "de": -0.2934570422649831, "fr": -0.027539061410728212, "nl": 0.036254883209861914, "es": -0.018505850177322414 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.009054632186889648, "delta_genuineness": 0.0067781019210815426, "delta_wer": 0.005462074157726332, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.10380371530848091, "ci95_z": [ -0.21363770182312192, -0.04147461072325172 ], "language_deltas": { "en": -0.06547852184385268, "de": -0.32099610367571124, "fr": -0.05771484842942268, "nl": -0.0397216762446231, "es": -0.03510742634879485 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.044081830978393556, "delta_genuineness": 0.0017573022842407226, "delta_wer": 0.0035059288537549406, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "emo_Pleasure_Ecstasy", "label": "Intense Pleasure Ecstasy", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Pleasure_Ecstasy", "adapter_sha256": "d955dc98b266853065b5348e6af4a5a26b4931135fa95324b73a65180a3cc854", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Pleasure_Ecstasy.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Pleasure_Ecstasy", "category": "emotion", "head": "emo_Pleasure_Ecstasy", "pole": "high", "label": "Intense Pleasure Ecstasy", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-1c5ef4ec2662", "general": "Project an overwhelming sense of blissful release, using vocal texture to convey expansive, uncontained joy.", "inline": [ "Begin with a breathy intake that immediately suggests an upward rush of sensation.", "Allow the mid-range of the voice to bloom with effortless, sustained resonance.", "Let the final articulation dissolve into a warm, audible exhalation of pure delight." ], "cfg": 3, "recipe": "P04", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.3608984345406619, "target_raw": 0.18399565428495407, "ce": 6.2631177234649655, "genuineness": 4.354964876174927, "identity": 0.5997853672504425, "wer": 0.0058095238095238096, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.5285815317777994, "target_raw": 0.7270501929521561, "ce": 6.267406778335571, "genuineness": 4.337140951156616, "identity": 0.5559286171197891, "wer": 0.01825098814229249, "valid_fraction": 0.98, "identity_delta": -0.043856750130653384, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.31316406277457776, "target_raw": 0.5955312845110893, "ce": 6.290758256912231, "genuineness": 4.35370964050293, "identity": 0.5600521922111511, "wer": 0.00823640127987954, "valid_fraction": 1, "identity_delta": -0.039733175039291385, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.2628234887240066, "target_raw": 0.564796834886074, "ce": 6.274078598022461, "genuineness": 4.385215377807617, "identity": 0.528790972828865, "wer": 0.011014492753623187, "valid_fraction": 1, "identity_delta": -0.07099439442157746, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.05034057405057112, "ci95_z": [ -0.04686033974421556, 0.16163207751635264 ], "language_deltas": { "en": -0.025683576588889968, "de": 0.0042724348335448515, "fr": -0.09360349018842135, "nl": 0.23356323098701554, "es": 0.13315427120960655 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.016679658889770507, "delta_genuineness": -0.0315057373046875, "delta_wer": -0.0027780914737436475, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.2154174690032217, "ci95_z": [ -0.2857397349409353, -0.16663084057405056 ], "language_deltas": { "en": -0.18159177975202576, "de": -0.2000976471736796, "fr": -0.14418943180708776, "nl": -0.20325928194864787, "es": -0.3479492043346676 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.023351478576660156, "delta_genuineness": 0.016568689346313475, "delta_wer": -0.010014586862412949, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Project an overwhelming sense of blissful release, using vocal texture to convey expansive, uncontained joy.\nSCRIPT:\n(Begin with a breathy intake that immediately suggests an upward rush of sensation.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Allow the mid-range of the voice to bloom with effortless, sustained resonance.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Let the final articulation dissolve into a warm, audible exhalation of pure delight.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "language": "fr", "seed": 1701, "scale": 0.0, "cfg": 2.0, "frames": 128, "reference": true } }, { "target": "emo_Pride", "label": "Intense Pride", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Pride", "adapter_sha256": "20be53fcac89d3ad7f8ee8656c1dc8bcc52945c00f5e1e082ae3ba2469c18d9f", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Pride.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Pride", "category": "emotion", "head": "emo_Pride", "pole": "high", "label": "Intense Pride", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-b8282fc33fc9", "general": "Focus on precise articulation and clarity, using controlled projection to make the pride feel earned and meticulous rather than boastful.", "inline": [ "Initiate with crisp, deliberate consonants, projecting a focused sense of personal victory.", "Sustain the mid-phrase with a clear, unwavering pitch contour that speaks to deep conviction.", "Conclude with a precise, resonant finality, letting the volume drop slightly while maintaining tonal authority." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 0.6489257902132568, "target_raw": 0.5577175208926201, "ce": 6.268836450576782, "genuineness": 4.341785192489624, "identity": 0.5516516341269017, "wer": 0.00981453634085213, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.2132421927350974, "target_raw": 0.8511124265193939, "ce": 6.2532688236236575, "genuineness": 4.150868048667908, "identity": 0.5154661053419113, "wer": 0.016608913588318622, "valid_fraction": 0.96, "identity_delta": -0.03618552878499031, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.6398242259801665, "target_raw": 1.0728975719213485, "ce": 6.249615211486816, "genuineness": 4.041288371086121, "identity": 0.5544497139751912, "wer": 0.0066012858232537865, "valid_fraction": 1, "identity_delta": 0.0027980798482894896, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.5525781293976608, "target_raw": 1.0275372737646102, "ce": 6.232459020614624, "genuineness": 4.058798542022705, "identity": 0.5445647957921028, "wer": 0.004552983248635422, "valid_fraction": 1, "identity_delta": -0.007086838334798813, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.08724609658250555, "ci95_z": [ -0.056757816436399086, 0.29349608038908237 ], "language_deltas": { "en": 0.4781249724138587, "de": 0.10898442072313348, "fr": -0.07587893601929782, "nl": -0.01796873826693341, "es": -0.05703123593823319 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.017156190872192383, "delta_genuineness": -0.01751017093658447, "delta_wer": 0.002048302574618364, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.42658203324506916, "ci95_z": [ 0.22996093057211678, 0.6064453160423113 ], "language_deltas": { "en": 0.4919921774015018, "de": 0.7253906455552578, "fr": 0.2596679551472558, "nl": 0.11113283227156068, "es": 0.5447265558497696 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.0036536121368408205, "delta_genuineness": -0.1095796775817871, "delta_wer": -0.010007627765064836, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Focus on precise articulation and clarity, using controlled projection to make the pride feel earned and meticulous rather than boastful.\nSCRIPT:\n(Initiate with crisp, deliberate consonants, projecting a focused sense of personal victory.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Sustain the mid-phrase with a clear, unwavering pitch contour that speaks to deep conviction.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Conclude with a precise, resonant finality, letting the volume drop slightly while maintaining tonal authority.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "language": "de", "seed": 1702, "scale": 0.0, "cfg": 2.0, "frames": 128, "reference": true } }, { "target": "emo_Sexual_Lust", "label": "Intense Sexual Lust", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Sexual_Lust", "adapter_sha256": "0bddaeb16329be0b4a16d501a61aad5945d817987bb1386db96eed90fadda87a", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Sexual_Lust.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Sexual_Lust", "category": "emotion", "head": "emo_Sexual_Lust", "pole": "high", "label": "Intense Sexual Lust", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.5, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-078e46d099b5", "general": "Use vocal texture to imply closeness, leaning into a slightly husky, textured quality that suggests suppressed yearning.", "inline": [ "Start with a slight vocal fry, but keep it very soft, suggesting a private, whispered confidence.", "Maintain a consistent, slightly lowered speaking register that feels inherently confidential and close.", "End with a very gentle vocal tremor, a barely perceptible quiver that speaks to deep feeling." ], "cfg": 2.5, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.49000000299379737, "target_raw": 0.010952902436256408, "ce": 6.247550077438355, "genuineness": 4.435423698425293, "identity": 0.6110569679737091, "wer": 0.01201072840203275, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.1823535151402621, "target_raw": 0.16407916009426116, "ce": 5.973448486328125, "genuineness": 4.522535133361816, "identity": 0.5611807900667191, "wer": 0.013698004893657068, "valid_fraction": 0.98, "identity_delta": -0.04987617790699005, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.14785644585834176, "target_raw": 0.18124954015016556, "ce": 5.814078722000122, "genuineness": 4.468059091567993, "identity": 0.5750786733627319, "wer": 0.01841709015622059, "valid_fraction": 1, "identity_delta": -0.03597829461097717, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.13186523876103012, "target_raw": 0.18920891493558883, "ce": 5.862568778991699, "genuineness": 4.470569496154785, "identity": 0.5411085352301598, "wer": 0.01441323169584039, "valid_fraction": 1, "identity_delta": -0.06994843274354935, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.01599120709731165, "ci95_z": [ -0.07365722528750485, 0.03099122056939991 ], "language_deltas": { "en": 0.048925792564449136, "de": -0.011572269340934064, "fr": 0.03281250439713991, "nl": -0.12548828386372546, "es": -0.024633779243487773 }, "languages_improved": 2, "clear_gain": false, "delta_ce": -0.04849005699157715, "delta_genuineness": -0.002510404586791992, "delta_wer": 0.004003858460380199, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.03449706928192032, "ci95_z": [ -0.07774414432690065, 0.14673828289074128 ], "language_deltas": { "en": 0.17631836209105248, "de": -0.11708984691336793, "fr": 0.03530274207291846, "nl": -0.09492188494034291, "es": 0.1728759740993415 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.15936976432800293, "delta_genuineness": -0.05447604179382324, "delta_wer": 0.004719085262563523, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Use vocal texture to imply closeness, leaning into a slightly husky, textured quality that suggests suppressed yearning.\nSCRIPT:\n(Start with a slight vocal fry, but keep it very soft, suggesting a private, whispered confidence.) [3.2 seconds duration] The road turns left behind the bridge.\n(Maintain a consistent, slightly lowered speaking register that feels inherently confidential and close.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(End with a very gentle vocal tremor, a barely perceptible quiver that speaks to deep feeling.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "language": "en", "seed": 1702, "scale": 0.0, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "emo_Shame", "label": "Intense Shame", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Shame", "adapter_sha256": "b9798e1bb370bf69f261b4eaa8161c257cdac1ce4d32a83892ccdb33b62d2be5", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Shame.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Shame", "category": "emotion", "head": "emo_Shame", "pole": "high", "label": "Intense Shame", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-48b6c1c43c18", "general": "Treat the vocal tract as constricted; the shame manifests as a physical tightening around the mouth and throat.", "inline": [ "Begin with a constricted resonance, as if the vocal cords are physically strained before speaking.", "Maintain a very tight, pressurized quality throughout, preventing any natural vocal release or expansion.", "Conclude with a sudden, sharp intake of breath that punctuates the silence immediately after the final utterance." ], "cfg": 3, "recipe": "P05", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.366015625559837, "target_raw": 0.019357388615608217, "ce": 6.249932880401611, "genuineness": 4.395006003379822, "identity": 0.5780217316746712, "wer": 0.012803500846979107, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.32818359492521343, "target_raw": 0.039182184040546415, "ce": 5.97718150138855, "genuineness": 4.3810731792449955, "identity": 0.599761166665703, "wer": 0.024749294184076794, "valid_fraction": 0.98, "identity_delta": 0.021739434991031883, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.33763671909878734, "target_raw": 0.034228544235229495, "ce": 5.8869527816772464, "genuineness": 4.318187265396118, "identity": 0.5977948135137559, "wer": 0.026372106154714848, "valid_fraction": 1, "identity_delta": 0.019773081839084626, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.3274999997156383, "target_raw": 0.039540402591228485, "ce": 5.862965936660767, "genuineness": 4.288062314987183, "identity": 0.5583475360274315, "wer": 0.02820816864295125, "valid_fraction": 1, "identity_delta": -0.019674195647239684, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.010136719383149024, "ci95_z": [ -0.02437500376779208, 0.004101565001494032 ], "language_deltas": { "en": 0.00878906461049674, "de": 0.009667972493354748, "fr": -0.02832031203346914, "nl": -0.016406249200232814, "es": -0.024414072785894646 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.02398684501647949, "delta_genuineness": 0.030124950408935546, "delta_wer": -0.0018360624882364015, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.009453124173573905, "ci95_z": [ -0.024882811713562263, 0.008417969223195588 ], "language_deltas": { "en": 0.02441405856781133, "de": -0.0022460903398815735, "fr": -0.01904296716157351, "nl": -0.016015621578773696, "es": -0.03437500035545208 }, "languages_improved": 1, "clear_gain": false, "delta_ce": -0.09022871971130371, "delta_genuineness": -0.06288591384887696, "delta_wer": 0.0016228119706380576, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Treat the vocal tract as constricted; the shame manifests as a physical tightening around the mouth and throat.\nSCRIPT:\n(Begin with a constricted resonance, as if the vocal cords are physically strained before speaking.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Maintain a very tight, pressurized quality throughout, preventing any natural vocal release or expansion.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Conclude with a sudden, sharp intake of breath that punctuates the silence immediately after the final utterance.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "language": "de", "seed": 1702, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "emo_Triumph", "label": "Intense Triumph", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/emo_Triumph", "adapter_sha256": "11d2616ed2ea09056a588a91482c234e542f9ad722641a0654d3f0b18d0d786c", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/emo_Triumph.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "emo_Triumph", "category": "emotion", "head": "emo_Triumph", "pole": "high", "label": "Intense Triumph", "floor": 3.0 }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-f9f56f50b9e3", "general": "Employ an accelerating, breathless quality, treating the triumph as an unstoppable, joyous momentum that carries the speech forward.", "inline": [ "Start with a rapid, energized delivery, implying the immediate aftermath of a massive effort.", "Increase the pace and vocal projection steadily, maintaining high forward momentum through the core of the thought.", "Let the energy peak on the last word, sustaining a bright, elated projection that feels like an outward surge." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.08333008132963059, "target_raw": 0.14911001771688462, "ce": 6.282656679153442, "genuineness": 4.3770565414428715, "identity": 0.577014462351799, "wer": 0.011366836062488237, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.627802758910763, "target_raw": 1.052615792155266, "ce": 6.095845069885254, "genuineness": 4.061936511993408, "identity": 0.5063609171658754, "wer": 0.00463579898362507, "valid_fraction": 1, "identity_delta": -0.07065354518592358, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.43159179523134533, "target_raw": 0.4209970845282078, "ce": 6.271378116607666, "genuineness": 4.325467443466186, "identity": 0.5488957750797272, "wer": 0.004718614718614718, "valid_fraction": 1, "identity_delta": -0.028118687272071837, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.3406250109356503, "target_raw": 0.37296515330672264, "ce": 6.2575578784942625, "genuineness": 4.281033139228821, "identity": 0.5709965917468071, "wer": 0.0047619047619047615, "valid_fraction": 1, "identity_delta": -0.006017870604991913, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.09096678429569503, "ci95_z": [ -0.039072271920715586, 0.2362597442993796 ], "language_deltas": { "en": 0.24062499259197884, "de": 0.31884762572498426, "fr": -0.07485351449505624, "nl": -0.015869148270033745, "es": -0.013916034073397953 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.01382023811340332, "delta_genuineness": 0.044434304237365725, "delta_wer": -4.329004329004321e-05, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -1.1962109636794176, "ci95_z": [ -1.549843798557815, -0.8305468968095671 ], "language_deltas": { "en": -0.5582031374789881, "de": -0.8950683533173994, "fr": -1.638378946894902, "nl": -1.3063476624123271, "es": -1.5830567182934718 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.1755330467224121, "delta_genuineness": 0.26353093147277834, "delta_wer": 8.2815734989648e-05, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "vn_S_CART_reg", "label": "Strong Cartoonish Style", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_CART_reg", "adapter_sha256": "37eaf41c9799bc1b7b1c996e15ce64fa8c1123cd876a03c45e9437fd70acc261", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_CART_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_CART_reg", "category": "talking_style", "head": "vn_S_CART_reg", "pole": "high", "label": "Strong Cartoonish Style", "floor": 4.0, "canonical_dimension_name": "Cartoonish Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.25, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-cff870ad7de8", "general": "Embrace a highly elastic vocal range, treating every word as a potential peak or valley in an exaggerated emotional landscape.", "inline": [ "Start with a sudden, bright upward lilt, making the initial sound feel slightly too bouncy for the context.", "Maintain rapid, almost frantic pitch shifts between syllables, creating a constant sense of playful instability.", "End the delivery with a sudden, slightly clipped high note that snaps the energy off abruptly." ], "cfg": 2.5, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-2.5", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.8384375023942975, "target_raw": 0.552873146533966, "ce": 6.266294784545899, "genuineness": 4.309651837348938, "identity": 0.5579182085394859, "wer": 0.01030679465462074, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.4728515638331883, "target_raw": 1.053430882692337, "ce": 6.206407051086426, "genuineness": 4.311409173011779, "identity": 0.5486446915566922, "wer": 0.008484848484848484, "valid_fraction": 1, "identity_delta": -0.009273516982793807, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.4494921895405944, "target_raw": 1.0854143762588502, "ce": 6.218956470489502, "genuineness": 4.306388335227966, "identity": 0.5586121982336044, "wer": 0.011129305477131564, "valid_fraction": 1, "identity_delta": 0.0006939896941184998, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.4728515638331883, "target_raw": 1.053430882692337, "ce": 6.206407051086426, "genuineness": 4.311409173011779, "identity": 0.5486446915566922, "wer": 0.008484848484848484, "valid_fraction": 1, "identity_delta": -0.009273516982793807, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.023359374292593944, "ci95_z": [ -0.023847654250217492, 0.07382812287778183 ], "language_deltas": { "en": -0.019140627448713286, "de": 0.10703124619089047, "fr": 0.06718750326495104, "nl": -0.04267576580950238, "es": 0.004394515265343871 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.012549419403076172, "delta_genuineness": -0.005020837783813476, "delta_wer": 0.002644456992283079, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.023359374292593944, "ci95_z": [ -0.023847654250217492, 0.07382812287778183 ], "language_deltas": { "en": -0.019140627448713286, "de": 0.10703124619089047, "fr": 0.06718750326495104, "nl": -0.04267576580950238, "es": 0.004394515265343871 }, "languages_improved": 3, "clear_gain": false, "delta_ce": 0.012549419403076172, "delta_genuineness": -0.005020837783813476, "delta_wer": 0.002644456992283079, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Embrace a highly elastic vocal range, treating every word as a potential peak or valley in an exaggerated emotional landscape.\nSCRIPT:\n(Start with a sudden, bright upward lilt, making the initial sound feel slightly too bouncy for the context.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Maintain rapid, almost frantic pitch shifts between syllables, creating a constant sense of playful instability.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(End the delivery with a sudden, slightly clipped high note that snaps the energy off abruptly.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "language": "fr", "seed": 1702, "scale": 0.0, "cfg": 1.25, "frames": 128, "reference": true } }, { "target": "vn_S_NEWS_reg", "label": "Strong Newsreader Style", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_NEWS_reg", "adapter_sha256": "9a42ad7d59768255765d74ccedbc42cbff52c1efac60bb20f9a2be94efb7ae98", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_NEWS_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_NEWS_reg", "category": "talking_style", "head": "vn_S_NEWS_reg", "pole": "high", "label": "Strong Newsreader Style", "floor": 4.0, "canonical_dimension_name": "Newsreader Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 4.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "manual-59e65fa10d8f", "general": "Deliver with a highly polished, almost academic neutrality, prioritizing impeccable articulation.", "inline": [ "Start with a crisp, clear articulation that establishes immediate factual distance.", "Ensure the pace remains consistently regulated, avoiding any natural speech ebb and flow.", "Preserve this highly controlled, precise delivery right up to the very last sound." ], "cfg": 3, "recipe": "P06", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": 1.1895898349333596, "target_raw": 1.1429302948713302, "ce": 6.2137142848968505, "genuineness": 4.343165912628174, "identity": 0.5770200660824776, "wer": 0.024086698960840838, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 2.082812522750439, "target_raw": 1.6351003074645996, "ce": 6.2736020755767825, "genuineness": 4.122876906394959, "identity": 0.5433347749710083, "wer": 0.011742449007895232, "valid_fraction": 0.98, "identity_delta": -0.03368529111146927, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 2.256601569806001, "target_raw": 1.7308589112758637, "ce": 6.239766302108765, "genuineness": 4.03871518611908, "identity": 0.5451197719573975, "wer": 0.016878460974570814, "valid_fraction": 1, "identity_delta": -0.03190029412508011, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 2.2890624747480266, "target_raw": 1.7487450265884399, "ce": 6.223245525360108, "genuineness": 4.057104024887085, "identity": 0.5571359089016914, "wer": 0.007527811623921464, "valid_fraction": 1, "identity_delta": -0.019884157180786132, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.032460904942025526, "ci95_z": [ -0.1382031033735609, 0.09523443945676088 ], "language_deltas": { "en": 0.012890576025707434, "de": -0.11660152220924247, "fr": 0.21562512879520596, "nl": -0.09062494219427764, "es": -0.18359376512752093 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.016520776748657227, "delta_genuineness": -0.018388838768005372, "delta_wer": 0.00935064935064935, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.17378904705556175, "ci95_z": [ 0.04460938259756497, 0.33832028875655307 ], "language_deltas": { "en": 0.24023428638991245, "de": 0.033398450747144745, "fr": 0.4726562355485694, "nl": 0.08906251369082896, "es": 0.03359374890135321 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.033835773468017576, "delta_genuineness": -0.08416172027587891, "delta_wer": 0.005136011966675582, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Deliver with a highly polished, almost academic neutrality, prioritizing impeccable articulation.\nSCRIPT:\n(Start with a crisp, clear articulation that establishes immediate factual distance.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Ensure the pace remains consistently regulated, avoiding any natural speech ebb and flow.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Preserve this highly controlled, precise delivery right up to the very last sound.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "language": "nl", "seed": 1702, "scale": 0.0, "cfg": 4.0, "frames": 128, "reference": true } }, { "target": "vn_S_RANT_reg", "label": "Strong Ranting/Angry Style", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_RANT_reg", "adapter_sha256": "e0115cd14bcb37c42dd9c7a79f7c14f5af0556895140514566fcc606273de7b4", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_RANT_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_RANT_reg", "category": "talking_style", "head": "vn_S_RANT_reg", "pole": "high", "label": "Strong Ranting/Angry Style", "floor": 4.0, "canonical_dimension_name": "Ranting/Angry Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-9663bdd37f37", "general": "A forceful sustained rant; emphatic stresses, rising insistence and frustrated breath without losing intelligibility. Establish it immediately. A forceful sustained rant; emphatic stresses, rising insistence and frustrated breath without losing intelligibility. Sustain the delivery. A forceful sustained rant; emphatic stresses, rising insistence and frustrated breath without losing intelligibility. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable.", "inline": [ "A forceful sustained rant; emphatic stresses, rising insistence and frustrated breath without losing intelligibility. Establish it immediately. Begin with the feeling held back, then let it leak through and become unmistakable.", "A forceful sustained rant; emphatic stresses, rising insistence and frustrated breath without losing intelligibility. Sustain the delivery. Begin with the feeling held back, then let it leak through and become unmistakable.", "A forceful sustained rant; emphatic stresses, rising insistence and frustrated breath without losing intelligibility. Preserve the intended style through the end. Begin with the feeling held back, then let it leak through and become unmistakable." ], "cfg": 3, "recipe": "P06", "strategy": "initial-physical-directing", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.9619531219708669, "target_raw": 0.2068141806125641, "ce": 6.235000667572021, "genuineness": 4.330864872932434, "identity": 0.5942860591411591, "wer": 0.012958780350084697, "valid_fraction": 0.98, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.598515642437588, "target_raw": 3.4738157987594604, "ce": 5.779726734161377, "genuineness": 3.631211919784546, "identity": 0.4897233945131302, "wer": 0.012799736495388668, "valid_fraction": 1, "identity_delta": -0.10456266462802886, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.2637304859010976, "target_raw": 3.046650404930115, "ce": 5.945490226745606, "genuineness": 3.687194199562073, "identity": 0.5347678869962692, "wer": 0.01113683418031244, "valid_fraction": 1, "identity_delta": -0.05951817214488983, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.17757813978071, "target_raw": 2.936725277900696, "ce": 5.947078790664673, "genuineness": 3.7133653020858763, "identity": 0.5075179040431976, "wer": 0.01296207415772633, "valid_fraction": 1, "identity_delta": -0.08676815509796143, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.0861523461203879, "ci95_z": [ -0.06556643241915547, 0.24556641186484493 ], "language_deltas": { "en": -0.07451168330366816, "de": 0.391406261459612, "fr": 0.11289067321066089, "nl": 0.1281249582490568, "es": -0.127148479013722 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.0015885639190673828, "delta_genuineness": -0.02617110252380371, "delta_wer": -0.0018252399774138905, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": -0.33478515653649027, "ci95_z": [ -0.488652320193928, -0.14453126689015433 ], "language_deltas": { "en": -0.4311523652732681, "de": 0.012499935767779635, "fr": -0.4652343177384353, "nl": -0.24921873532877695, "es": -0.5408203001097507 }, "languages_improved": 1, "clear_gain": false, "delta_ce": 0.1657634925842285, "delta_genuineness": 0.05598227977752686, "delta_wer": -0.001662902315076228, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "vn_S_TECH_reg", "label": "Strong teacher-like, didactic delivery", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_TECH_reg", "adapter_sha256": "c8af52a34d5af840c957333d13a5f45df8d7a99cf35628bd90a7348876a912ea", "recipe_status": "corrected_taxonomy_new_search", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_TECH_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_TECH_reg", "category": "talking_style", "head": "vn_S_TECH_reg", "pole": "high", "label": "Strong teacher-like, didactic delivery", "floor": 4.0, "canonical_dimension_name": "Teacher/Didactic Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 2.5, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P01", "general": "A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear.", "inline": [ "A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear.", "A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear.", "A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear." ], "recipe": "P01", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": 1.6541406254205706, "target_raw": 2.5433186626434328, "ce": 6.239766292572021, "genuineness": 4.251912288665771, "identity": 0.5765025797486305, "wer": 0.007556960583276372, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 1.7628125044962828, "target_raw": 2.6756738662719726, "ce": 6.2589875507354735, "genuineness": 4.142207112312317, "identity": 0.572125246822834, "wer": 0.012138647012788889, "valid_fraction": 0.98, "identity_delta": -0.004377332925796509, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 1.7932812620971417, "target_raw": 2.7127828097343443, "ce": 6.2575578784942625, "genuineness": 4.168315415382385, "identity": 0.5612384041026235, "wer": 0.00872362724994304, "valid_fraction": 1, "identity_delta": -0.015264175646007062, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 1.8271874788644125, "target_raw": 2.7540783524513244, "ce": 6.249932889938354, "genuineness": 4.123881039619445, "identity": 0.5622481724619866, "wer": 0.005647186147186147, "valid_fraction": 1, "identity_delta": -0.014254407286643981, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": -0.03390621676727083, "ci95_z": [ -0.07578120801940103, 0.007968774484859352 ], "language_deltas": { "en": -0.02734372193646656, "de": 0.025781261852445926, "fr": 0.007812535327935733, "nl": -0.08124994066130278, "es": -0.0945312184189665 }, "languages_improved": 2, "clear_gain": false, "delta_ce": 0.007624988555908203, "delta_genuineness": 0.04443437576293945, "delta_wer": 0.003076441102756892, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.03046875760085891, "ci95_z": [ -0.03968750015293475, 0.08421875674442411 ], "language_deltas": { "en": 0.11015626949918535, "de": 0.06953126980505482, "fr": -0.09765626582875035, "nl": 0.025000026916522478, "es": 0.04531248761228226 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.0014296722412109374, "delta_genuineness": 0.02610830307006836, "delta_wer": -0.0034150197628458495, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear.\nSCRIPT:\n(A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] Escribí el número en el sobre.\n(A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(A patient teacher guiding one learner, with clear conceptual grouping, explanatory emphasis and approachable pedagogical authority. Make this audible from the opening words and keep every word clear.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "language": "es", "seed": 1701, "scale": 0.0, "cfg": 2.5, "frames": 128, "reference": true } }, { "target": "vn_S_WHIS_reg", "label": "Strong Whisper-Talk Style", "status": "NO_QUALIFYING_NEW_RECIPE", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_S_WHIS_reg", "adapter_sha256": "efaee1aef8d7f5f4b4aa4e7be8925982f27333286d381083a110dad22ff72eee", "recipe_status": "direct_manual_recipe", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_S_WHIS_reg.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_S_WHIS_reg", "category": "talking_style", "head": "vn_S_WHIS_reg", "pole": "high", "label": "Strong Whisper-Talk Style", "floor": 4.0, "canonical_dimension_name": "Whisper-Talk Style" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "candidate": { "id": "manual-c79d4bcf72d9", "general": "Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.", "inline": [ "Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.", "Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.", "Fade the final words out slowly, letting the slight tremor in the voice remain present until silence." ], "cfg": 3, "recipe": "P04", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "heldout": { "standard": { "n": 50, "target_z": -0.6456640627708922, "target_raw": 0.7347310471534729, "ce": 6.291711368560791, "genuineness": 4.390236177444458, "identity": 0.5940428373217582, "wer": 0.009192076039902127, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.7236328242240977, "target_raw": 2.9237736463546753, "ce": 5.524906034469605, "genuineness": 4.4301518535614015, "identity": 0.5163092762231827, "wer": 0.02796586327478776, "valid_fraction": 0.98, "identity_delta": -0.0777335610985756, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.8048046794810085, "target_raw": 3.0535399889945984, "ce": 5.401293201446533, "genuineness": 4.389985132217407, "identity": 0.47779575005173686, "wer": 0.024120672233944544, "valid_fraction": 1, "identity_delta": -0.11624708727002143, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.7891015720686113, "target_raw": 3.0284360313415526, "ce": 5.397038917541504, "genuineness": 4.405800771713257, "identity": 0.5218156635761261, "wer": 0.03252567188722795, "valid_fraction": 0.96, "identity_delta": -0.07222717374563217, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.015703107412397138, "ci95_z": [ -0.028164086073446497, 0.04960936641883437 ], "language_deltas": { "en": 0.029492175076825732, "de": -0.06660159350696132, "fr": 0.06718749056246551, "nl": 0.037890616989746946, "es": 0.01054684793990882 }, "languages_improved": 4, "clear_gain": false, "delta_ce": 0.0042542839050292965, "delta_genuineness": -0.01581563949584961, "delta_wer": -0.008404999653283406, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.08117185525691085, "ci95_z": [ 0.052265602588768376, 0.10554686397206618 ], "language_deltas": { "en": 0.06425779649405801, "de": 0.09804687310666749, "fr": 0.12109374807753927, "nl": 0.08945307749191737, "es": 0.03300778111437212 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.12361283302307129, "delta_genuineness": -0.04016672134399414, "delta_wer": -0.0038451910408432146, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": null }, { "target": "vn_TENS_reg_low", "label": "very relaxed and unstrained", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_TENS_reg_low", "adapter_sha256": "f6ed187f358572964ec9903478ec361d219af9393b4596dd8cc10f38eb31097f", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_TENS_reg_low.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_TENS_reg_low", "category": "bipolar", "head": "vn_TENS_reg", "pole": "low", "label": "very relaxed and unstrained", "floor": 1.0, "canonical_dimension_name": "Tension" }, "configuration": { "adapter": true, "scale": 1.0, "cfg": 5.0, "reason": "Held-out target improves against matched CFG and best manual-only configuration; QA passes" }, "candidate": { "id": "canonical-extension-P02", "general": "Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "inline": [ "Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.", "Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality." ], "recipe": "P02", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": -0.2193212821470641, "target_raw": 1.8419780135154724, "ce": 6.243419923782349, "genuineness": 4.401030983924866, "identity": 0.5999285724759101, "wer": 0.011057312252964427, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.0640234448406569, "target_raw": 1.7127268528938293, "ce": 6.1396885681152344, "genuineness": 4.521907558441162, "identity": 0.6108118721842766, "wer": 0.016721249764728024, "valid_fraction": 1, "identity_delta": 0.010883299708366394, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.03870116350486073, "target_raw": 1.627231297492981, "ce": 6.107679615020752, "genuineness": 4.533455410003662, "identity": 0.6112429520487785, "wer": 0.013146740368708332, "valid_fraction": 1, "identity_delta": 0.011314379572868348, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.0012207175825132788, "target_raw": 1.6584254932403564, "ce": 6.097512950897217, "genuineness": 4.590567426681519, "identity": 0.6102218699455261, "wer": 0.014564276303406738, "valid_fraction": 0.98, "identity_delta": 0.010293297469615936, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.037480445922347456, "ci95_z": [ 0.009824196638509104, 0.06809568034210756 ], "language_deltas": { "en": -0.001416037781251002, "de": 0.09570310172474643, "fr": 0.046289036539140245, "nl": 0.03974606913194333, "es": 0.0070800599971582775 }, "languages_improved": 4, "clear_gain": true, "delta_ce": 0.010166664123535157, "delta_genuineness": -0.057112016677856446, "delta_wer": -0.0014175359346984058, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.10272460834551764, "ci95_z": [ 0.07494140938320722, 0.12732908945489213 ], "language_deltas": { "en": 0.08935550321527934, "de": 0.10461425814820079, "fr": 0.12363280444318146, "nl": 0.05390624641919176, "es": 0.14211422950173483 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.03200895309448242, "delta_genuineness": 0.0115478515625, "delta_wer": -0.003574509396019693, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.\nSCRIPT:\n(Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Audibly relaxed and unstrained delivery, with loose, comfortable voicing and easy articulation. Use breath, pitch movement, consonant attack and release to realize the requested quality.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "language": "nl", "seed": 1701, "scale": 1.0, "cfg": 5.0, "frames": 128, "reference": true } }, { "target": "vn_TEMP_reg_low", "label": "very slow speaking tempo", "status": "VALIDATED_HELDOUT", "reference_required": true, "languages": [ "en", "de", "fr", "nl", "es" ], "tested_seeds": [ 1701, 1702 ], "scope": "one fixed synthetic reference voice per language;unseen texts;not a no-reference or universal optimum claim", "model_repo": "laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs", "model_revision": "ed3fe8924e6164e2548058c07b144ae8ab76a308", "base_subfolder": "base_model", "rank4_already_merged": true, "adapter_subfolder": "adapters/vn_TEMP_reg_low", "adapter_sha256": "f9c6904318158a5102c75b265ad4cae131211c72b8f49d9e4696492b4e30f67b", "recipe_status": "new_canonical_dimension_extension", "comparison_url": "https://laion-humaneness-voice-large-intense-bucket-loras.static.hf.space/vn_TEMP_reg_low.html#manual-comparison", "original_manual_url": "https://laion-humaneness-voice-large-prompting-manuals.static.hf.space/models/rank4/index.html", "definition": { "id": "vn_TEMP_reg_low", "category": "bipolar", "head": "vn_TEMP_reg", "pole": "low", "label": "very slow speaking tempo", "floor": 1.0, "canonical_dimension_name": "Tempo" }, "configuration": { "adapter": false, "scale": 0.0, "cfg": 3.0, "reason": "Use manual-only configuration; no clear qualifying incremental adapter gain" }, "candidate": { "id": "canonical-extension-P04", "general": "Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis.", "inline": [ "Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis.", "Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis.", "Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis." ], "recipe": "P04", "cfg": 1.0, "timing_syntax": "canonical", "canonical_labels": true, "lexical_proxy": false, "origin": "new_canonical_manual_extension_requires_new_evidence" }, "heldout": { "standard": { "n": 50, "target_z": -0.23684569796869895, "target_raw": 2.3993338441848753, "ce": 6.238654289245606, "genuineness": 4.344170107841491, "identity": 0.6038601872324943, "wer": 0.022059466848940532, "valid_fraction": 0.96, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": -0.0627636615643371, "target_raw": 2.1903204250335695, "ce": 6.216097116470337, "genuineness": 4.364629979133606, "identity": 0.589395119547844, "wer": 0.01012987012987013, "valid_fraction": 0.98, "identity_delta": -0.014465067684650421, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": -0.05624512230598687, "target_raw": 2.182493872642517, "ce": 6.199973487854004, "genuineness": 4.325593028068543, "identity": 0.6106709486246109, "wer": 0.0073738942217203085, "valid_fraction": 1, "identity_delta": 0.006810761392116547, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": -0.0627636615643371, "target_raw": 2.1903204250335695, "ce": 6.216097116470337, "genuineness": 4.364629979133606, "identity": 0.589395119547844, "wer": 0.01012987012987013, "valid_fraction": 0.98, "identity_delta": -0.014465067684650421, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.006518539258350239, "ci95_z": [ -0.019887707049435703, 0.03282225644294456 ], "language_deltas": { "en": 0.023681600338368147, "de": -0.02014163209016796, "fr": -0.0343261678317929, "nl": 0.05336918566264259, "es": 0.010009710212701316 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.016123628616333006, "delta_genuineness": -0.039036951065063476, "delta_wer": -0.002755975908149821, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.006518539258350239, "ci95_z": [ -0.019887707049435703, 0.03282225644294456 ], "language_deltas": { "en": 0.023681600338368147, "de": -0.02014163209016796, "fr": -0.0343261678317929, "nl": 0.05336918566264259, "es": 0.010009710212701316 }, "languages_improved": 3, "clear_gain": false, "delta_ce": -0.016123628616333006, "delta_genuineness": -0.039036951065063476, "delta_wer": -0.002755975908149821, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "example": { "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis.\nSCRIPT:\n(Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis.) [3.2 seconds duration] The road turns left behind the bridge.\n(Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Very slow speaking tempo, with deliberately elongated syllables and spacious articulation of each word. Think while speaking, with small organic hesitations and spontaneous emphasis.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "language": "en", "seed": 1701, "scale": 0.0, "cfg": 3.0, "frames": 128, "reference": true } } ], "native_cfg": "code/bucket_native_cfg.py", "compiler": "code/bucket_tool_infer.py", "reference_contract": "63–187MOSS-v2frames,5.04–14.96seconds", "best_tested_only": true, "no_reference_validated": false, "selection_uses_heldout": false }