Fabio
c36510c21e
Some checks failed
Linux Build / Linux cpu (push) Has been cancelled
Linux Build / Linux cpu-clang (push) Has been cancelled
Linux Build / Linux vulkan (push) Has been cancelled
mac-build / macOS CPU (push) Has been cancelled
Nix Build / Nix (cpu) (push) Has been cancelled
Nix Build / Nix (vulkan) (push) Has been cancelled
windows-build / Windows CPU (push) Has been cancelled
347 lines
9.5 KiB
JSON
347 lines
9.5 KiB
JSON
{
|
|
"family": "chatterbox",
|
|
"display_name": "Chatterbox",
|
|
"description": "Open-source Chatterbox family for expressive TTS and voice conversion, with emotion exaggeration control, fast generation, zero-shot voice cloning, and an integrated multilingual TTS path.",
|
|
"category": "tts",
|
|
"status": "supported",
|
|
"tasks": [
|
|
"tts",
|
|
"clone",
|
|
"vc"
|
|
],
|
|
"modes": [
|
|
"offline"
|
|
],
|
|
"languages": [
|
|
"ar",
|
|
"da",
|
|
"de",
|
|
"el",
|
|
"en",
|
|
"es",
|
|
"fi",
|
|
"fr",
|
|
"hi",
|
|
"it",
|
|
"ko",
|
|
"ms",
|
|
"nl",
|
|
"no",
|
|
"pl",
|
|
"pt",
|
|
"sv",
|
|
"sw",
|
|
"tr"
|
|
],
|
|
"options": {
|
|
"request": [
|
|
{
|
|
"name": "exaggeration",
|
|
"type": "float",
|
|
"description": "Emotion exaggeration. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": 0.5
|
|
},
|
|
{
|
|
"name": "guidance_scale",
|
|
"type": "float",
|
|
"description": "T3 classifier-free guidance scale. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": 0.5
|
|
},
|
|
{
|
|
"name": "temperature",
|
|
"type": "float",
|
|
"description": "T3 sampling temperature. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": 0.8
|
|
},
|
|
{
|
|
"name": "repetition_penalty",
|
|
"type": "float",
|
|
"description": "T3 repetition penalty. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": 1.2
|
|
},
|
|
{
|
|
"name": "min_p",
|
|
"type": "float",
|
|
"description": "Minimum sampling probability. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": 0.05
|
|
},
|
|
{
|
|
"name": "top_p",
|
|
"type": "float",
|
|
"description": "Nucleus sampling probability. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": 1
|
|
},
|
|
{
|
|
"name": "flow_guidance_scale",
|
|
"type": "float",
|
|
"description": "S3Gen flow guidance scale. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": 0.7
|
|
},
|
|
{
|
|
"name": "max_tokens",
|
|
"type": "int",
|
|
"description": "Maximum generated speech tokens. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 384
|
|
},
|
|
{
|
|
"name": "do_sample",
|
|
"type": "bool",
|
|
"description": "Enable sampling. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": true
|
|
},
|
|
{
|
|
"name": "stop_on_eos",
|
|
"type": "bool",
|
|
"description": "Stop token generation on EOS. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"default": true
|
|
},
|
|
{
|
|
"name": "num_inference_steps",
|
|
"type": "int",
|
|
"description": "Voice conversion flow steps; max_steps is also accepted.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 10
|
|
},
|
|
{
|
|
"name": "seed",
|
|
"type": "int",
|
|
"description": "Sampling seed; omitted chooses a random seed. Captured during clone preparation; changing it on the same prepared session is rejected.",
|
|
"required": false,
|
|
"min": 0,
|
|
"max": 4294967295
|
|
},
|
|
{
|
|
"name": "source_audio",
|
|
"type": "audio_path",
|
|
"description": "Voice conversion source WAV path; overrides audio_input.",
|
|
"required": false
|
|
},
|
|
{
|
|
"name": "target_voice",
|
|
"type": "audio_path",
|
|
"description": "Voice conversion target WAV path; overrides the speaker reference.",
|
|
"required": false
|
|
},
|
|
{
|
|
"name": "text_chunk_size",
|
|
"type": "int",
|
|
"description": "Framework long-form text chunk size.",
|
|
"required": false,
|
|
"default": 128
|
|
}
|
|
],
|
|
"session": [
|
|
{
|
|
"name": "weight_type",
|
|
"type": "enum",
|
|
"description": "Weight storage type; defaults to native.",
|
|
"required": false,
|
|
"preset": "weight_type_full"
|
|
},
|
|
{
|
|
"name": "t3_weight_type",
|
|
"type": "enum",
|
|
"description": "T3 weight storage; inherits weight_type.",
|
|
"required": false,
|
|
"preset": "weight_type_full"
|
|
},
|
|
{
|
|
"name": "multilingual_t3",
|
|
"type": "enum",
|
|
"description": "Multilingual T3 checkpoint selection.",
|
|
"required": false,
|
|
"values": [
|
|
"v2",
|
|
"v3"
|
|
],
|
|
"default": "v2"
|
|
},
|
|
{
|
|
"name": "conditionals_cache_slots",
|
|
"type": "int",
|
|
"description": "Cached speaker conditionals; zero disables reuse.",
|
|
"required": false,
|
|
"min": 0,
|
|
"default": 1
|
|
},
|
|
{
|
|
"name": "mem_saver",
|
|
"type": "bool",
|
|
"description": "Release staged runtime components.",
|
|
"required": false,
|
|
"default": false
|
|
},
|
|
{
|
|
"name": "encoder_condition_samples",
|
|
"type": "int",
|
|
"description": "Maximum reference samples for the T3 tokenizer; the voice encoder still consumes the full 16 kHz reference.",
|
|
"required": false,
|
|
"default": 96000
|
|
},
|
|
{
|
|
"name": "decoder_condition_samples",
|
|
"type": "int",
|
|
"description": "S3Gen reference sample limit at 24 kHz.",
|
|
"required": false,
|
|
"default": 240000
|
|
},
|
|
{
|
|
"name": "t3_speech_cond_prompt_len",
|
|
"type": "int",
|
|
"description": "T3 reference speech prompt token limit.",
|
|
"required": false,
|
|
"default": 150
|
|
}
|
|
],
|
|
"load": []
|
|
},
|
|
"capabilities": {
|
|
"clone": [
|
|
"speaker_reference"
|
|
],
|
|
"vc": [
|
|
"speaker_reference"
|
|
]
|
|
},
|
|
"runtime": {
|
|
"tags": [
|
|
"gguf"
|
|
]
|
|
},
|
|
"ui": {
|
|
"recommended_package": "chatterbox_q8_0",
|
|
"tags": [
|
|
"TTS",
|
|
"Clone",
|
|
"VC",
|
|
"GGUF"
|
|
],
|
|
"docs": [
|
|
"docs/tts.md",
|
|
"docs/gguf.md"
|
|
]
|
|
},
|
|
"package_defaults": {
|
|
"download": {
|
|
"kind": "huggingface_snapshot",
|
|
"repo": "audio-cpp/audio.cpp-gguf",
|
|
"revision": "main",
|
|
"gated": false
|
|
}
|
|
},
|
|
"packages": [
|
|
{
|
|
"id": "chatterbox_q8_0",
|
|
"display_name": "Chatterbox Q8_0 GGUF",
|
|
"default": true,
|
|
"format": "gguf",
|
|
"precision": "q8_0",
|
|
"target_directory": "Chatterbox-GGUF",
|
|
"files": [
|
|
"Chatterbox-GGUF/chatterbox-q8_0.gguf"
|
|
],
|
|
"strip_prefix": "Chatterbox-GGUF"
|
|
},
|
|
{
|
|
"id": "chatterbox_f16",
|
|
"display_name": "Chatterbox F16 GGUF",
|
|
"format": "gguf",
|
|
"precision": "f16",
|
|
"target_directory": "Chatterbox-GGUF",
|
|
"files": [
|
|
"Chatterbox-GGUF/chatterbox-f16.gguf"
|
|
],
|
|
"strip_prefix": "Chatterbox-GGUF"
|
|
},
|
|
{
|
|
"id": "chatterbox_safetensors",
|
|
"display_name": "Chatterbox Safetensors",
|
|
"format": "safetensors",
|
|
"precision": "native",
|
|
"target_directory": "chatterbox",
|
|
"files": [
|
|
"ve.safetensors",
|
|
"t3_cfg.safetensors",
|
|
"s3gen.safetensors",
|
|
"tokenizer.json",
|
|
"grapheme_mtl_merged_expanded_v1.json",
|
|
"Cangjie5_TC.json",
|
|
"conds.pt",
|
|
"t3_mtl23ls_v2.safetensors",
|
|
"t3_mtl23ls_v3.safetensors"
|
|
],
|
|
"download": {
|
|
"kind": "huggingface_snapshot",
|
|
"repo": "ResembleAI/chatterbox"
|
|
}
|
|
}
|
|
],
|
|
"sources": [
|
|
{
|
|
"format": "gguf",
|
|
"roots": {
|
|
"model": ".",
|
|
"weights": "$gguf"
|
|
},
|
|
"files": {
|
|
"english_tokenizer": "model:tokenizer.json",
|
|
"multilingual_tokenizer": "model:grapheme_mtl_merged_expanded_v1.json",
|
|
"cangjie_mapping": "model:Cangjie5_TC.json",
|
|
"builtin_conditionals": "model:conds.pt"
|
|
},
|
|
"tensors": {
|
|
"voice_encoder_weights": {
|
|
"source": "weights:",
|
|
"prefix": "voice_encoder"
|
|
},
|
|
"s3gen_weights": {
|
|
"source": "weights:",
|
|
"prefix": "s3gen"
|
|
},
|
|
"t3_english_weights": {
|
|
"source": "weights:",
|
|
"prefix": "t3_english"
|
|
},
|
|
"t3_multilingual_v2_weights": {
|
|
"source": "weights:",
|
|
"prefix": "t3_multilingual_v2"
|
|
},
|
|
"t3_multilingual_v3_weights": {
|
|
"source": "weights:",
|
|
"prefix": "t3_multilingual_v3"
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"format": "safetensors",
|
|
"roots": {
|
|
"model": "."
|
|
},
|
|
"files": {
|
|
"english_tokenizer": "model:tokenizer.json",
|
|
"multilingual_tokenizer": "model:grapheme_mtl_merged_expanded_v1.json",
|
|
"cangjie_mapping": "model:Cangjie5_TC.json",
|
|
"builtin_conditionals": "model:conds.pt"
|
|
},
|
|
"tensors": {
|
|
"voice_encoder_weights": "model:ve.safetensors",
|
|
"s3gen_weights": "model:s3gen.safetensors",
|
|
"t3_english_weights": "model:t3_cfg.safetensors",
|
|
"t3_multilingual_v2_weights": "model:t3_mtl23ls_v2.safetensors",
|
|
"t3_multilingual_v3_weights": "model:t3_mtl23ls_v3.safetensors"
|
|
}
|
|
}
|
|
]
|
|
}
|