Instructions to use second-state/OuteTTS-0.2-500M-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use second-state/OuteTTS-0.2-500M-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="second-state/OuteTTS-0.2-500M-GGUF",
	filename="OuteTTS-0.2-500M-Q2_K.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use second-state/OuteTTS-0.2-500M-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf second-state/OuteTTS-0.2-500M-GGUF: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 second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf second-state/OuteTTS-0.2-500M-GGUF: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 second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M

Use Docker

docker model run hf.co/second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use second-state/OuteTTS-0.2-500M-GGUF with Ollama:
```
ollama run hf.co/second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M
```

Unsloth Studio

How to use second-state/OuteTTS-0.2-500M-GGUF 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 second-state/OuteTTS-0.2-500M-GGUF 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 second-state/OuteTTS-0.2-500M-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for second-state/OuteTTS-0.2-500M-GGUF to start chatting

Atomic Chat new
Docker Model Runner
How to use second-state/OuteTTS-0.2-500M-GGUF with Docker Model Runner:
```
docker model run hf.co/second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M
```

Lemonade

How to use second-state/OuteTTS-0.2-500M-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull second-state/OuteTTS-0.2-500M-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.OuteTTS-0.2-500M-GGUF-Q4_K_M

List all available models

lemonade list

OuteTTS-0.2-500M-GGUF

Original Model

OuteAI/OuteTTS-0.2-500M

Run with LlamaEdge

LlamaEdge version: v0.14.9

Run as LlamaEdge service

wasmedge --dir .:. \
  --nn-preload tts:GGML:AUTO:OuteTTS-0.2-500M-Q5_K_M.gguf \
  llama-api-server.wasm config \
  --file llama_server_config.toml \
  --tts

llama_server_config.toml can be derived from the template config file llama_server_config.toml.bkp. The recommended [tts] config is shown as below

[tts]
model_name   = "tts"            # Name of the TTS model.
model_alias  = "tts"            # Alias of the TTS model.
codec_model  = ""               # Required. Path to the codec model file.
speaker_file  = ""              # Path to an alternative speaker file.
ctx_size     = 8192             # Context size. Default is 8192.
batch_size   = 8192             # Batch size. Default is 8192.
ubatch_size  = 8192             # Physical maximum batch size. Default is 8192.
n_predict    = 4096             # Number of tokens to predict. Default is 4096.
n_gpu_layers = 100              # Number of layers to run on GPU. Default is 100.
temp         = 0.8              # Temperature. Default is 0.8.

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
OuteTTS-0.2-500M-Q2_K.gguf	Q2_K	2	344 MB	smallest, significant quality loss - not recommended for most purposes
OuteTTS-0.2-500M-Q3_K_L.gguf	Q3_K_L	3	375 MB	small, substantial quality loss
OuteTTS-0.2-500M-Q3_K_M.gguf	Q3_K_M	3	361 MB	very small, high quality loss
OuteTTS-0.2-500M-Q3_K_S.gguf	Q3_K_S	3	344 MB	very small, high quality loss
OuteTTS-0.2-500M-Q4_0.gguf	Q4_0	4	358 MB	legacy; small, very high quality loss - prefer using Q3_K_M
OuteTTS-0.2-500M-Q4_K_M.gguf	Q4_K_M	4	403 MB	medium, balanced quality - recommended
OuteTTS-0.2-500M-Q4_K_S.gguf	Q4_K_S	4	391 MB	small, greater quality loss
OuteTTS-0.2-500M-Q5_0.gguf	Q5_0	5	402 MB	legacy; medium, balanced quality - prefer using Q4_K_M
OuteTTS-0.2-500M-Q5_K_M.gguf	Q5_K_M	5	426 MB	large, very low quality loss - recommended
OuteTTS-0.2-500M-Q5_K_S.gguf	Q5_K_S	5	418 MB	large, low quality loss - recommended
OuteTTS-0.2-500M-Q6_K.gguf	Q6_K	6	511 MB	very large, extremely low quality loss
OuteTTS-0.2-500M-Q8_0.gguf	Q8_0	8	537 MB	very large, extremely low quality loss - not recommended
OuteTTS-0.2-500M-f16.gguf	f16	16	1.00 GB
wavtokenizer-large-75-ggml-f16.gguf	f16	16	1.00 GB

Quantized with llama.cpp b4381

Downloads last month: 410

GGUF

Model size

0.5B params

Architecture

qwen2

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Model tree for second-state/OuteTTS-0.2-500M-GGUF

Base model

OuteAI/OuteTTS-0.2-500M

Quantized

(9)

this model