Instructions to use AgeOfAlgorithms/Llasa-3b-GPTQ-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use AgeOfAlgorithms/Llasa-3b-GPTQ-4bit with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="AgeOfAlgorithms/Llasa-3b-GPTQ-4bit")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("AgeOfAlgorithms/Llasa-3b-GPTQ-4bit")
model = AutoModelForCausalLM.from_pretrained("AgeOfAlgorithms/Llasa-3b-GPTQ-4bit")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use AgeOfAlgorithms/Llasa-3b-GPTQ-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "AgeOfAlgorithms/Llasa-3b-GPTQ-4bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AgeOfAlgorithms/Llasa-3b-GPTQ-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/AgeOfAlgorithms/Llasa-3b-GPTQ-4bit

SGLang

How to use AgeOfAlgorithms/Llasa-3b-GPTQ-4bit with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "AgeOfAlgorithms/Llasa-3b-GPTQ-4bit" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AgeOfAlgorithms/Llasa-3b-GPTQ-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "AgeOfAlgorithms/Llasa-3b-GPTQ-4bit" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AgeOfAlgorithms/Llasa-3b-GPTQ-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use AgeOfAlgorithms/Llasa-3b-GPTQ-4bit with Docker Model Runner:
```
docker model run hf.co/AgeOfAlgorithms/Llasa-3b-GPTQ-4bit
```

Configuration Parsing Warning:Config file tokenizer_config.json cannot be fetched (too big)

Model Description

This is a 4bit GPTQ quantization of Llasa-3B by the HKUSTAudio team. I tested using a script written by GitHub user nivibilla, linked below. For some reason, I was not able to run it on my RTX 3090, while quantized Llasa-1B worked fine. Please let me know if you can get it working.

Model Sources

Repository: HKUSTAudio/Llasa-3B
Paper: LLaSA: Scaling Train-Time and Inference-Time Compute for LLaMA-based Speech Synthesis (Coming soon)
Test Script: https://github.com/slives-lab/local-llasa-tts_voice/blob/main/llasa_vllm_longtext_inference.ipynb

Downloads last month: 2

Safetensors

Model size

3B params

Tensor type

I32

F16

Model tree for AgeOfAlgorithms/Llasa-3b-GPTQ-4bit

Base model

meta-llama/Llama-3.2-3B-Instruct

Finetuned

HKUSTAudio/Llasa-3B

Quantized

(8)

this model