Hugging Face's logo Hugging Face
  • Models
  • Datasets
  • Spaces
  • Buckets new
  • Docs
  • Enterprise
  • Pricing

  • Log In
  • Sign Up

1bitLLM
/
bitnet_b1_58-3B

Text Generation
Transformers
Safetensors
llama
text-generation-inference
Model card Files Files and versions
xet
Community
12

Instructions to use 1bitLLM/bitnet_b1_58-3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

  • Libraries
  • Transformers

    How to use 1bitLLM/bitnet_b1_58-3B with Transformers:

    # Use a pipeline as a high-level helper
    from transformers import pipeline
    
    pipe = pipeline("text-generation", model="1bitLLM/bitnet_b1_58-3B")
    # Load model directly
    from transformers import AutoTokenizer, AutoModelForCausalLM
    
    tokenizer = AutoTokenizer.from_pretrained("1bitLLM/bitnet_b1_58-3B")
    model = AutoModelForCausalLM.from_pretrained("1bitLLM/bitnet_b1_58-3B")
  • Notebooks
  • Google Colab
  • Kaggle
  • Local Apps
  • vLLM

    How to use 1bitLLM/bitnet_b1_58-3B with vLLM:

    Install from pip and serve model
    # Install vLLM from pip:
    pip install vllm
    # Start the vLLM server:
    vllm serve "1bitLLM/bitnet_b1_58-3B"
    # Call the server using curl (OpenAI-compatible API):
    curl -X POST "http://localhost:8000/v1/completions" \
    	-H "Content-Type: application/json" \
    	--data '{
    		"model": "1bitLLM/bitnet_b1_58-3B",
    		"prompt": "Once upon a time,",
    		"max_tokens": 512,
    		"temperature": 0.5
    	}'
    Use Docker
    docker model run hf.co/1bitLLM/bitnet_b1_58-3B
  • SGLang

    How to use 1bitLLM/bitnet_b1_58-3B 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 "1bitLLM/bitnet_b1_58-3B" \
        --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": "1bitLLM/bitnet_b1_58-3B",
    		"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 "1bitLLM/bitnet_b1_58-3B" \
            --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": "1bitLLM/bitnet_b1_58-3B",
    		"prompt": "Once upon a time,",
    		"max_tokens": 512,
    		"temperature": 0.5
    	}'
  • Docker Model Runner

    How to use 1bitLLM/bitnet_b1_58-3B with Docker Model Runner:

    docker model run hf.co/1bitLLM/bitnet_b1_58-3B
New discussion
Resources
  • PR & discussions documentation
  • Code of Conduct
  • Hub documentation

Add metadata and link to paper

#12 opened about 1 year ago by
nielsr

can you provide wikitest ppl and c4 ppl separately?

#11 opened almost 2 years ago by
sheropen-2

Can you provide more details on the training?

1
#10 opened almost 2 years ago by
dequ777

Any plans to use MQA (multi-query attention) or GQA (grouped-query attention) in the future?

#9 opened about 2 years ago by
graefics

Efficient Inference Kernel Support for 1.58bit.

❤️ 10
#8 opened about 2 years ago by
LeiWang1999

This code from BitLinear doesn't make sense

1
#7 opened about 2 years ago by
qmsoqm

Is it bitnet {-1,0,1}?

4
#6 opened about 2 years ago by
Remek

ValueError: Tokenizer class BitnetTokenizer does not exist or is not currently imported.

👍 2
4
#5 opened about 2 years ago by
ryanzhangofficial

Longer inference time

2
#4 opened about 2 years ago by
dittops

Why are these models fp32?

👀 3
5
#2 opened about 2 years ago by
supercharge19

Is there a chat/instruct model in plans?

2
#1 opened about 2 years ago by
MrVodnik
Company
TOS Privacy About Careers
Website
Models Datasets Spaces Pricing Docs