Instructions to use tsumeone/llama-30b-supercot-3bit-128g-cuda with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsumeone/llama-30b-supercot-3bit-128g-cuda with Transformers:

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

pipe = pipeline("text-generation", model="tsumeone/llama-30b-supercot-3bit-128g-cuda")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tsumeone/llama-30b-supercot-3bit-128g-cuda")
model = AutoModelForCausalLM.from_pretrained("tsumeone/llama-30b-supercot-3bit-128g-cuda")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use tsumeone/llama-30b-supercot-3bit-128g-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-3bit-128g-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-3bit-128g-cuda",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tsumeone/llama-30b-supercot-3bit-128g-cuda

SGLang

How to use tsumeone/llama-30b-supercot-3bit-128g-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-3bit-128g-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-3bit-128g-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-3bit-128g-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-3bit-128g-cuda",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tsumeone/llama-30b-supercot-3bit-128g-cuda with Docker Model Runner:
```
docker model run hf.co/tsumeone/llama-30b-supercot-3bit-128g-cuda
```

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

3bit quantized version of this: https://huggingface.co/ausboss/llama-30b-supercot

GPTQ quantization using https://github.com/0cc4m/GPTQ-for-LLaMa

Made at the request of someone that wanted a 3bit version. The file is 17% smaller than 4bit non-groupsize, but the wikitext2 ppl is 12% worse. I don't have a functioning Ooba install so I can't test this myself.

Command used to quantize:
python llama.py c:\llama-30b-supercot c4 --wbits 3 --true-sequential --groupsize 128 --save_safetensors 4bit-128g.safetensors

Evaluation & Score (Lower is better):

WikiText2: 5.22 (12% worse than 4bit non-groupsize)
PTB: 19.63 (11% worse than 4bit non-groupsize)
C4: 6.93 (7% worse than 4bit non-groupsize)

4bit non-groupsize version is here: https://huggingface.co/tsumeone/llama-30b-supercot-4bit-cuda

4bit 128 groupsize version is here: https://huggingface.co/tsumeone/llama-30b-supercot-4bit-128g-cuda

Downloads last month: 1,027