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
350 lines
8.9 KiB
JSON
350 lines
8.9 KiB
JSON
{
|
|
"family": "moss_tts_nano",
|
|
"display_name": "MOSS-TTS-Nano",
|
|
"description": "Compact deployment-first MOSS-TTS model for real-time multilingual speech generation, lightweight integration, and zero-shot voice cloning.",
|
|
"category": "tts",
|
|
"status": "supported",
|
|
"tasks": [
|
|
"tts",
|
|
"clone"
|
|
],
|
|
"modes": [
|
|
"offline",
|
|
"streaming"
|
|
],
|
|
"languages": [
|
|
"ar",
|
|
"cs",
|
|
"da",
|
|
"de",
|
|
"el",
|
|
"en",
|
|
"es",
|
|
"fa",
|
|
"fr",
|
|
"hu",
|
|
"it",
|
|
"ja",
|
|
"ko",
|
|
"pl",
|
|
"pt",
|
|
"ru",
|
|
"sv",
|
|
"tr",
|
|
"zh"
|
|
],
|
|
"options": {
|
|
"request": [
|
|
{
|
|
"name": "reference_text",
|
|
"type": "string",
|
|
"description": "Transcript for reference audio; a nonempty value requires voice reference audio.",
|
|
"required": false,
|
|
"default": ""
|
|
},
|
|
{
|
|
"name": "max_tokens",
|
|
"type": "int",
|
|
"description": "Maximum generated audio frames.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 300
|
|
},
|
|
{
|
|
"name": "active_codebooks",
|
|
"type": "int",
|
|
"description": "Active codebooks; defaults to the checkpoint n_vq. Must not exceed checkpoint n_vq; omitted uses n_vq.",
|
|
"required": false,
|
|
"min": 1
|
|
},
|
|
{
|
|
"name": "do_sample",
|
|
"type": "bool",
|
|
"description": "Enable sampling.",
|
|
"required": false,
|
|
"default": true
|
|
},
|
|
{
|
|
"name": "temperature",
|
|
"type": "float",
|
|
"description": "Audio sampling temperature. Must be positive, including when do_sample=false.",
|
|
"required": false,
|
|
"default": 1.7
|
|
},
|
|
{
|
|
"name": "top_p",
|
|
"type": "float",
|
|
"description": "Audio nucleus sampling probability.",
|
|
"required": false,
|
|
"default": 0.8,
|
|
"min": 0,
|
|
"max": 1
|
|
},
|
|
{
|
|
"name": "top_k",
|
|
"type": "int",
|
|
"description": "Audio top-k sampling limit.",
|
|
"required": false,
|
|
"default": 25,
|
|
"min": 0
|
|
},
|
|
{
|
|
"name": "repetition_penalty",
|
|
"type": "float",
|
|
"description": "Audio repetition penalty. Must be positive.",
|
|
"required": false,
|
|
"default": 1
|
|
},
|
|
{
|
|
"name": "text_temperature",
|
|
"type": "float",
|
|
"description": "Text sampling temperature. Must be positive, including when do_sample=false.",
|
|
"required": false,
|
|
"default": 1.5
|
|
},
|
|
{
|
|
"name": "text_top_p",
|
|
"type": "float",
|
|
"description": "Text nucleus sampling probability.",
|
|
"required": false,
|
|
"default": 1,
|
|
"min": 0,
|
|
"max": 1
|
|
},
|
|
{
|
|
"name": "text_top_k",
|
|
"type": "int",
|
|
"description": "Text top-k sampling limit.",
|
|
"required": false,
|
|
"default": 50,
|
|
"min": 0
|
|
},
|
|
{
|
|
"name": "seed",
|
|
"type": "int",
|
|
"description": "Sampling seed; omitted chooses a random seed.",
|
|
"required": false,
|
|
"min": 0,
|
|
"max": 4294967295
|
|
},
|
|
{
|
|
"name": "text_chunk_size",
|
|
"type": "int",
|
|
"description": "Framework long-form text chunk size; default 50 when hanzi, kana and Hangul outnumber other letters and digits, otherwise 256.",
|
|
"required": false,
|
|
"min": 1
|
|
},
|
|
{
|
|
"name": "text_chunk_mode",
|
|
"type": "enum",
|
|
"description": "Framework text chunking mode.",
|
|
"required": false,
|
|
"values": [
|
|
"default",
|
|
"tag_aware",
|
|
"japanese",
|
|
"endline"
|
|
],
|
|
"default": "default"
|
|
}
|
|
],
|
|
"session": [
|
|
{
|
|
"name": "weight_type",
|
|
"type": "enum",
|
|
"description": "Weight storage type.",
|
|
"required": false,
|
|
"preset": "weight_type_full",
|
|
"default": "native"
|
|
},
|
|
{
|
|
"name": "global_weight_type",
|
|
"type": "enum",
|
|
"description": "Global transformer weight storage; inherits weight_type.",
|
|
"required": false,
|
|
"preset": "weight_type_full"
|
|
},
|
|
{
|
|
"name": "local_frame_weight_type",
|
|
"type": "enum",
|
|
"description": "Local frame decoder weight storage; inherits weight_type.",
|
|
"required": false,
|
|
"preset": "weight_type_full"
|
|
},
|
|
{
|
|
"name": "prefill_mode",
|
|
"type": "enum",
|
|
"description": "How prompt rows are forwarded: one batched graph (default, fast) or one row at a time through the decode graph. single_row keeps quantized runs on a single kernel family for numerical self-consistency at a one-time prefill latency cost; see the model docs' quantization notes.",
|
|
"required": false,
|
|
"values": [
|
|
"batched",
|
|
"single_row"
|
|
],
|
|
"default": "batched"
|
|
},
|
|
{
|
|
"name": "global_prefill_graph_arena_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 256
|
|
},
|
|
{
|
|
"name": "global_decode_graph_arena_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 128
|
|
},
|
|
{
|
|
"name": "global_weight_context_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 512
|
|
},
|
|
{
|
|
"name": "local_frame_graph_arena_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 64
|
|
},
|
|
{
|
|
"name": "local_frame_weight_context_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 128
|
|
},
|
|
{
|
|
"name": "audio_tokenizer_encoder_graph_arena_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 64
|
|
},
|
|
{
|
|
"name": "audio_tokenizer_decoder_graph_arena_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 64
|
|
},
|
|
{
|
|
"name": "audio_tokenizer_weight_context_mb",
|
|
"type": "int",
|
|
"description": "Context or graph arena size in MiB.",
|
|
"required": false,
|
|
"min": 1,
|
|
"default": 128
|
|
}
|
|
],
|
|
"load": []
|
|
},
|
|
"capabilities": {
|
|
"clone": [
|
|
"speaker_reference"
|
|
]
|
|
},
|
|
"runtime": {
|
|
"tags": [
|
|
"gguf"
|
|
]
|
|
},
|
|
"ui": {
|
|
"recommended_package": "moss_tts_nano_100m_q8_0",
|
|
"tags": [
|
|
"TTS",
|
|
"Clone",
|
|
"GGUF"
|
|
],
|
|
"docs": [
|
|
"docs/models/moss_tts_nano.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": "moss_tts_nano_100m_q8_0",
|
|
"display_name": "MOSS-TTS-Nano 100M Q8_0 GGUF",
|
|
"default": true,
|
|
"format": "gguf",
|
|
"precision": "q8_0",
|
|
"target_directory": "MOSS-TTS-Nano-100M-GGUF",
|
|
"files": [
|
|
"MOSS-TTS-Nano-100M-GGUF/moss-tts-nano-100m-q8_0.gguf"
|
|
],
|
|
"strip_prefix": "MOSS-TTS-Nano-100M-GGUF"
|
|
},
|
|
{
|
|
"id": "moss_tts_nano_100m_bf16",
|
|
"display_name": "MOSS-TTS-Nano 100M BF16 GGUF",
|
|
"format": "gguf",
|
|
"precision": "bf16",
|
|
"target_directory": "MOSS-TTS-Nano-100M-GGUF",
|
|
"files": [
|
|
"MOSS-TTS-Nano-100M-GGUF/moss-tts-nano-100m-bf16.gguf"
|
|
],
|
|
"strip_prefix": "MOSS-TTS-Nano-100M-GGUF"
|
|
}
|
|
],
|
|
"sources": [
|
|
{
|
|
"format": "gguf",
|
|
"roots": {
|
|
"model": ".",
|
|
"weights": "$gguf"
|
|
},
|
|
"files": {
|
|
"config": "model:config.json",
|
|
"tokenizer_config": "model:tokenizer_config.json",
|
|
"tokenizer_model": "model:tokenizer.model",
|
|
"audio_tokenizer_config": "model:audio_tokenizer/config.json"
|
|
},
|
|
"tensors": {
|
|
"model_weights": {
|
|
"source": "weights:",
|
|
"prefix": "model_weights"
|
|
},
|
|
"audio_tokenizer_weights": {
|
|
"source": "weights:",
|
|
"prefix": "audio_tokenizer_weights"
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"format": "safetensors",
|
|
"roots": {
|
|
"model": ".",
|
|
"audio_tokenizer": "audio_tokenizer"
|
|
},
|
|
"files": {
|
|
"config": "model:config.json",
|
|
"tokenizer_config": "model:tokenizer_config.json",
|
|
"tokenizer_model": "model:tokenizer.model",
|
|
"audio_tokenizer_config": "audio_tokenizer:config.json"
|
|
},
|
|
"tensors": {
|
|
"model_weights": "model:model.safetensors",
|
|
"audio_tokenizer_weights": "audio_tokenizer:model.safetensors.index.json"
|
|
}
|
|
}
|
|
]
|
|
}
|