Instructions to use TheBloke/starchat-beta-GPTQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TheBloke/starchat-beta-GPTQ with Transformers:

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

pipe = pipeline("text-generation", model="TheBloke/starchat-beta-GPTQ")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TheBloke/starchat-beta-GPTQ")
model = AutoModelForCausalLM.from_pretrained("TheBloke/starchat-beta-GPTQ", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use TheBloke/starchat-beta-GPTQ with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/TheBloke/starchat-beta-GPTQ

SGLang

How to use TheBloke/starchat-beta-GPTQ 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 "TheBloke/starchat-beta-GPTQ" \
    --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": "TheBloke/starchat-beta-GPTQ",
		"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 "TheBloke/starchat-beta-GPTQ" \
        --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": "TheBloke/starchat-beta-GPTQ",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use TheBloke/starchat-beta-GPTQ with Docker Model Runner:
```
docker model run hf.co/TheBloke/starchat-beta-GPTQ
```

Sequence Length?

by gsaivinay - opened Jun 25, 2023

Discussion

gsaivinay

Jun 25, 2023

Hello,
Thanks for providing this. Awesome work.

The original model sequence length is 8192. Is the same length is used while converting to GPTQ?

gsaivinay

Jul 1, 2023

@TheBloke - would you be able to answer this?

TheBloke

Owner Jul 8, 2023

Sorry for the delay. Yes it's 8K in GPTQ also

gsaivinay

Jul 8, 2023

Thank you very much for your reply. Would you be able to provide the data set name used for the conversion? I'd like to replicate with different setting.

TheBloke

Owner Jul 8, 2023

For the GPTQ? I used wikitext2

gsaivinay changed discussion status to closed Jul 8, 2023

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment