Instructions to use GreenBitAI/LLaMA-3B-2bit-groupsize32 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use GreenBitAI/LLaMA-3B-2bit-groupsize32 with Transformers:

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

pipe = pipeline("text-generation", model="GreenBitAI/LLaMA-3B-2bit-groupsize32")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("GreenBitAI/LLaMA-3B-2bit-groupsize32")
model = AutoModelForCausalLM.from_pretrained("GreenBitAI/LLaMA-3B-2bit-groupsize32")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use GreenBitAI/LLaMA-3B-2bit-groupsize32 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "GreenBitAI/LLaMA-3B-2bit-groupsize32"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "GreenBitAI/LLaMA-3B-2bit-groupsize32",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/GreenBitAI/LLaMA-3B-2bit-groupsize32

SGLang

How to use GreenBitAI/LLaMA-3B-2bit-groupsize32 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 "GreenBitAI/LLaMA-3B-2bit-groupsize32" \
    --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": "GreenBitAI/LLaMA-3B-2bit-groupsize32",
		"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 "GreenBitAI/LLaMA-3B-2bit-groupsize32" \
        --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": "GreenBitAI/LLaMA-3B-2bit-groupsize32",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use GreenBitAI/LLaMA-3B-2bit-groupsize32 with Docker Model Runner:
```
docker model run hf.co/GreenBitAI/LLaMA-3B-2bit-groupsize32
```

large 2bit models

by KnutJaegersberg - opened Oct 6, 2023

Discussion

KnutJaegersberg

Oct 6, 2023

It would be great to have 2 bit versions of some larger models, like

https://huggingface.co/CofeAI/FLM-101B

or galactica 120b for using their work token for reasoning. and fine tuned falcon-180b or bloomchat and vulture-180b.

https://huggingface.co/sambanovasystems/BLOOMChat-176B-v1

https://huggingface.co/vilm/vulture-180b

https://huggingface.co/alpayariyak/LIMA-180b

NicoNico

GREENBITAI org Oct 6, 2023

The request has been received, and I believe the larger version will soon be available.

KnutJaegersberg changed discussion status to closed Oct 8, 2023

KnutJaegersberg

Oct 8, 2023

There is also a large llama-70b fine tune with increased context length. Being able to use that more thanks to 2 bit quantization would be a practical combo, too.
https://huggingface.co/abacusai/Giraffe-v2-70b-32k

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