Instructions to use niobures/MOSS-TTS with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use niobures/MOSS-TTS with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="niobures/MOSS-TTS",
	filename="models/MOSS-TTS-GGUF/MOSS_TTS_F16.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use niobures/MOSS-TTS with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf niobures/MOSS-TTS:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf niobures/MOSS-TTS:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf niobures/MOSS-TTS:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf niobures/MOSS-TTS:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf niobures/MOSS-TTS:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf niobures/MOSS-TTS:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf niobures/MOSS-TTS:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf niobures/MOSS-TTS:Q4_K_M

Use Docker

docker model run hf.co/niobures/MOSS-TTS:Q4_K_M

LM Studio
Jan
Ollama
How to use niobures/MOSS-TTS with Ollama:
```
ollama run hf.co/niobures/MOSS-TTS:Q4_K_M
```

Unsloth Studio

How to use niobures/MOSS-TTS with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for niobures/MOSS-TTS to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for niobures/MOSS-TTS to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for niobures/MOSS-TTS to start chatting

Docker Model Runner
How to use niobures/MOSS-TTS with Docker Model Runner:
```
docker model run hf.co/niobures/MOSS-TTS:Q4_K_M
```

Lemonade

How to use niobures/MOSS-TTS with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull niobures/MOSS-TTS:Q4_K_M

Run and chat with the model

lemonade run user.MOSS-TTS-Q4_K_M

List all available models

lemonade list

MOSS-TTS / models /MOSS-TTS-Local-Transformer /config.json

niobures

MOSS-TTS-Local-Transformer

5d5fac5 verified about 1 month ago

raw

history blame contribute delete

2.21 kB

	{
	"model_type": "moss_tts_delay",
	"architectures": [
	"MossTTSDelayModel"
	],
	"auto_map": {
	"AutoConfig": "configuration_moss_tts.MossTTSDelayConfig",
	"AutoModel": "modeling_moss_tts.MossTTSDelayModel"
	},
	"dtype": "bfloat16",
	"initializer_range": 0.02,
	"language_config": {
	"_name_or_path": "Qwen/Qwen3-8B",
	"architectures": [
	"Qwen3ForCausalLM"
	],
	"attention_bias": false,
	"attention_dropout": 0.0,
	"bos_token_id": 151643,
	"eos_token_id": 151645,
	"pad_token_id": 151643,
	"head_dim": 128,
	"hidden_act": "silu",
	"hidden_size": 2048,
	"initializer_range": 0.02,
	"intermediate_size": 6144,
	"layer_types": [
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention"
	],
	"max_position_embeddings": 40960,
	"max_window_layers": 28,
	"model_type": "qwen3",
	"num_attention_heads": 16,
	"num_hidden_layers": 28,
	"num_key_value_heads": 8,
	"rms_norm_eps": 1e-06,
	"rope_scaling": null,
	"rope_theta": 1000000,
	"sliding_window": null,
	"use_cache": true,
	"use_sliding_window": false,
	"vocab_size": 155648
	},
	"n_vq": 32,
	"audio_vocab_size": 1024,
	"audio_user_slot_token_id": 151654,
	"audio_assistant_gen_slot_token_id": 151656,
	"audio_assistant_delay_slot_token_id": 151662,
	"audio_start_token_id": 151652,
	"audio_end_token_id": 151653,
	"audio_pad_code": 1024,
	"sampling_rate": 24000,
	"transformers_version": "4.57.1",

	"additional_mlp_ffn_hidden_size": 2048,
	"local_ffn_hidden_size": 8960,
	"local_hidden_size": 1536,
	"local_num_layers": 4
	}