158 lines
3.8 KiB
JSON
158 lines
3.8 KiB
JSON
{
|
|
"family": "omnivoice",
|
|
"display_name": "OmniVoice",
|
|
"description": "Massively multilingual zero-shot TTS model from k2-fsa for 600+ languages, supporting short-reference voice cloning, attribute-based voice design, pronunciation controls, and nonverbal tags.",
|
|
"category": "tts",
|
|
"status": "supported",
|
|
"tasks": [
|
|
"tts",
|
|
"clone",
|
|
"design"
|
|
],
|
|
"modes": [
|
|
"offline",
|
|
"streaming"
|
|
],
|
|
"languages": [
|
|
"600+ languages"
|
|
],
|
|
"capabilities": {
|
|
"clone": [
|
|
"speaker_reference"
|
|
],
|
|
"design": [
|
|
"voice_design"
|
|
]
|
|
},
|
|
"runtime": {
|
|
"tags": [
|
|
"gguf",
|
|
"stream"
|
|
]
|
|
},
|
|
"ui": {
|
|
"recommended_package": "omnivoice_q8_0",
|
|
"tags": [
|
|
"TTS",
|
|
"Clone",
|
|
"Design",
|
|
"GGUF",
|
|
"Stream"
|
|
],
|
|
"docs": [
|
|
"docs/models/omnivoice.md",
|
|
"docs/tts.md",
|
|
"docs/gguf.md"
|
|
]
|
|
},
|
|
"package_defaults": {
|
|
"download": {
|
|
"kind": "huggingface_snapshot",
|
|
"repo": "audio-cpp/audio.cpp-gguf",
|
|
"revision": "main",
|
|
"gated": false
|
|
}
|
|
},
|
|
"packages": [
|
|
{
|
|
"id": "omnivoice_q8_0",
|
|
"display_name": "OmniVoice Q8_0 GGUF",
|
|
"default": true,
|
|
"format": "gguf",
|
|
"precision": "q8_0",
|
|
"target_directory": "OmniVoice-GGUF",
|
|
"files": [
|
|
"OmniVoice-GGUF/omnivoice-q8_0.gguf"
|
|
],
|
|
"strip_prefix": "OmniVoice-GGUF"
|
|
},
|
|
{
|
|
"id": "omnivoice_bf16",
|
|
"display_name": "OmniVoice BF16 GGUF",
|
|
"format": "gguf",
|
|
"precision": "bf16",
|
|
"target_directory": "OmniVoice-GGUF",
|
|
"files": [
|
|
"OmniVoice-GGUF/omnivoice-bf16.gguf"
|
|
],
|
|
"strip_prefix": "OmniVoice-GGUF"
|
|
},
|
|
{
|
|
"id": "omnivoice_f16",
|
|
"display_name": "OmniVoice F16 GGUF",
|
|
"format": "gguf",
|
|
"precision": "f16",
|
|
"target_directory": "OmniVoice-GGUF",
|
|
"files": [
|
|
"OmniVoice-GGUF/omnivoice-f16.gguf"
|
|
],
|
|
"strip_prefix": "OmniVoice-GGUF"
|
|
},
|
|
{
|
|
"id": "omnivoice_safetensors",
|
|
"display_name": "OmniVoice Safetensors",
|
|
"format": "safetensors",
|
|
"precision": "native",
|
|
"target_directory": "OmniVoice",
|
|
"files": [
|
|
"config.json",
|
|
"model.safetensors",
|
|
"tokenizer.json"
|
|
],
|
|
"download": {
|
|
"kind": "huggingface_snapshot",
|
|
"repo": "k2-fsa/OmniVoice"
|
|
}
|
|
}
|
|
],
|
|
"sources": [
|
|
{
|
|
"format": "gguf",
|
|
"roots": {
|
|
"model": ".",
|
|
"weights": "$gguf"
|
|
},
|
|
"files": {
|
|
"config": "model:config.json",
|
|
"tokenizer_json": "model:tokenizer.json",
|
|
"tokenizer_config": "model:tokenizer_config.json",
|
|
"audio_tokenizer_config": "model:audio_tokenizer/config.json",
|
|
"audio_tokenizer_preprocessor": "model:audio_tokenizer/preprocessor_config.json"
|
|
},
|
|
"optional_files": {
|
|
"chat_template": "model:chat_template.jinja"
|
|
},
|
|
"tensors": {
|
|
"weights": {
|
|
"source": "weights:",
|
|
"prefix": "weights"
|
|
},
|
|
"audio_tokenizer_weights": {
|
|
"source": "weights:",
|
|
"prefix": "audio_tokenizer_weights"
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"format": "safetensors",
|
|
"roots": {
|
|
"model": "."
|
|
},
|
|
"files": {
|
|
"config": "model:config.json",
|
|
"tokenizer_json": "model:tokenizer.json",
|
|
"tokenizer_config": "model:tokenizer_config.json",
|
|
"audio_tokenizer_config": "model:audio_tokenizer/config.json",
|
|
"audio_tokenizer_preprocessor": "model:audio_tokenizer/preprocessor_config.json"
|
|
},
|
|
"optional_files": {
|
|
"chat_template": "model:chat_template.jinja"
|
|
},
|
|
"tensors": {
|
|
"weights": "model:model.safetensors",
|
|
"audio_tokenizer_weights": "model:audio_tokenizer/model.safetensors"
|
|
}
|
|
}
|
|
]
|
|
}
|