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

Libraries

How to use 1bitLLM/bitnet_b1_58-large with Transformers:

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

pipe = pipeline("text-generation", model="1bitLLM/bitnet_b1_58-large")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("1bitLLM/bitnet_b1_58-large")
model = AutoModelForCausalLM.from_pretrained("1bitLLM/bitnet_b1_58-large")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use 1bitLLM/bitnet_b1_58-large 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-large"
# 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-large",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/1bitLLM/bitnet_b1_58-large

SGLang

How to use 1bitLLM/bitnet_b1_58-large 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-large" \
    --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-large",
		"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-large" \
        --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-large",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use 1bitLLM/bitnet_b1_58-large with Docker Model Runner:
```
docker model run hf.co/1bitLLM/bitnet_b1_58-large
```

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

by qmsoqm - opened Apr 10, 2024

Discussion

qmsoqm

Apr 10, 2024

I'm getting this error when I try to run the model.

My code:

'''

Import necessary libraries

from transformers import AutoTokenizer, AutoModelWithLMHead, pipeline

Specify the model name

model_name = "1bitLLM/bitnet_b1_58-large"

Load the tokenizer and the model

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelWithLMHead.from_pretrained(model_name)

Create the pipeline

pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

Define the prompt

prompt = """
The weather in South Korea is like
"""

Generate the text

print(pipe(prompt))
'''

I'm using transformers 4.39.3.

I've found this github issue which seems relevant to this issue.

gyuryeol

Oct 30, 2024

I changed the tokenizer_class in tokenizer_config.json and the problem was resolved. but

ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating

The following problem occurred. Was there a solution to it?

chancar

Dec 3, 2024

Try:

from transformers import PreTrainedTokenizerFast
tokenizer = PreTrainedTokenizerFast(tokenizer_file="path_to/tokenizer.json")

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