Instructions to use tsumeone/llama-30b-supercot-4bit-cuda with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsumeone/llama-30b-supercot-4bit-cuda with Transformers:

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

pipe = pipeline("text-generation", model="tsumeone/llama-30b-supercot-4bit-cuda", device_map="auto")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tsumeone/llama-30b-supercot-4bit-cuda")
model = AutoModelForCausalLM.from_pretrained("tsumeone/llama-30b-supercot-4bit-cuda", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tsumeone/llama-30b-supercot-4bit-cuda with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tsumeone/llama-30b-supercot-4bit-cuda"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tsumeone/llama-30b-supercot-4bit-cuda",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tsumeone/llama-30b-supercot-4bit-cuda

SGLang

How to use tsumeone/llama-30b-supercot-4bit-cuda 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 "tsumeone/llama-30b-supercot-4bit-cuda" \
    --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": "tsumeone/llama-30b-supercot-4bit-cuda",
		"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 "tsumeone/llama-30b-supercot-4bit-cuda" \
        --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": "tsumeone/llama-30b-supercot-4bit-cuda",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tsumeone/llama-30b-supercot-4bit-cuda with Docker Model Runner:
```
docker model run hf.co/tsumeone/llama-30b-supercot-4bit-cuda
```

Instructions on how to load this?

by unluckyton - opened May 6, 2023

Discussion

unluckyton

May 6, 2023

Is this supposed to work on a 4090? New to add this and need some help.

I downloaded this into my koboldai's models folder. I tried it load it as is but got errors. I found https://huggingface.co/ausboss/llama-30b-supercot/tree/main and copied their pytorch_model-00001-of-00243.bin files into the directory, tried loading again. It started but ran out of VRAM at ~30%. Help please? Thanks

tsumeone

Owner May 6, 2023

You don't need those bin files. You do need the Occam latestgptq branch of koboldAI for 4bit model support. https://github.com/0cc4m/KoboldAI

unluckyton

May 7, 2023

Thanks, got it to work with that, appreciate the assistance, if I want to use textgen webui is the main repo fine or is there a special repo I should try for that too.

tsumeone

Owner May 7, 2023

I don't use ooba, but I think the main repo is fine with that one

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