Instructions to use hongyin/chat-llama2-70b-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hongyin/chat-llama2-70b-4bit with Transformers:

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

pipe = pipeline("text-generation", model="hongyin/chat-llama2-70b-4bit")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("hongyin/chat-llama2-70b-4bit")
model = AutoModelForCausalLM.from_pretrained("hongyin/chat-llama2-70b-4bit", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use hongyin/chat-llama2-70b-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "hongyin/chat-llama2-70b-4bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hongyin/chat-llama2-70b-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/hongyin/chat-llama2-70b-4bit

SGLang

How to use hongyin/chat-llama2-70b-4bit 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 "hongyin/chat-llama2-70b-4bit" \
    --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": "hongyin/chat-llama2-70b-4bit",
		"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 "hongyin/chat-llama2-70b-4bit" \
        --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": "hongyin/chat-llama2-70b-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use hongyin/chat-llama2-70b-4bit with Docker Model Runner:
```
docker model run hf.co/hongyin/chat-llama2-70b-4bit
```

hongyin/chat-llama2-70b-4bit

I am pleased to introduce a 4-bit quantized version of the Llama2-chat-70B based model, dedicated to reducing the cost of training and inference. Losing fat is the only way to solve all problems.

Human: Say "I love you" differently in Chinese.
Assistant: "I love you" in Chinese is "我爱你" (wǒ ài nǐ).
Human: How do you write "I love you" in Chinese characters?
Assistant: "I love you" in Chinese characters is 我爱你 (wǒ ài nǐ).
Human: What is the meaning of "我爱你"?
Assistant: "我爱你" (wǒ ài n

Bibtex entry and citation info

Please cite if you find it helpful.

@article{zhu2023metaaid,
  title={MetaAID 2.0: An Extensible Framework for Developing Metaverse Applications via Human-controllable Pre-trained Models},
  author={Zhu, Hongyin},
  journal={arXiv preprint arXiv:2302.13173},
  year={2023}
}

license: other

Downloads last month: 10

Paper for hongyin/chat-llama2-70b-4bit

MetaAID 2.0: An Extensible Framework for Developing Metaverse Applications via Human-controllable Pre-trained Models

Paper • 2302.13173 • Published Feb 25, 2023