Instructions to use VMware/open-llama-7b-open-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use VMware/open-llama-7b-open-instruct with Transformers:

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

pipe = pipeline("text-generation", model="VMware/open-llama-7b-open-instruct")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("VMware/open-llama-7b-open-instruct")
model = AutoModelForCausalLM.from_pretrained("VMware/open-llama-7b-open-instruct")

Inference
Local Apps Settings

vLLM

How to use VMware/open-llama-7b-open-instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "VMware/open-llama-7b-open-instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "VMware/open-llama-7b-open-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/VMware/open-llama-7b-open-instruct

SGLang

How to use VMware/open-llama-7b-open-instruct 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 "VMware/open-llama-7b-open-instruct" \
    --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": "VMware/open-llama-7b-open-instruct",
		"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 "VMware/open-llama-7b-open-instruct" \
        --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": "VMware/open-llama-7b-open-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use VMware/open-llama-7b-open-instruct with Docker Model Runner:
```
docker model run hf.co/VMware/open-llama-7b-open-instruct
```

AWQ 4-bit 128g version of open-llama-7b-open-instruct!

by abhinavkulkarni - opened Jul 5, 2023

Discussion

abhinavkulkarni

Jul 5, 2023

Hi,

I would like to draw everyone's attention to AWQ quantized version of open-llama-7b-open-instruct model at https://huggingface.co/abhinavkulkarni/open-llama-7b-open-instruct-w4-g128-awq.

For more on AWQ, click here.

The quantized model size on disk is 3.89GB vs 13.48GB for the original model. Similar gains could be observed for VRAM usage. The perplexity is only worse by 5%.

Please take a look and give it a try.

Thanks!

abhinavkulkarni changed discussion status to closed Jul 20, 2023

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