Instructions to use upstage/SOLAR-0-70b-16bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use upstage/SOLAR-0-70b-16bit with Transformers:

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

pipe = pipeline("text-generation", model="upstage/SOLAR-0-70b-16bit")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("upstage/SOLAR-0-70b-16bit")
model = AutoModelForCausalLM.from_pretrained("upstage/SOLAR-0-70b-16bit", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use upstage/SOLAR-0-70b-16bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "upstage/SOLAR-0-70b-16bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "upstage/SOLAR-0-70b-16bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/upstage/SOLAR-0-70b-16bit

SGLang

How to use upstage/SOLAR-0-70b-16bit 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 "upstage/SOLAR-0-70b-16bit" \
    --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": "upstage/SOLAR-0-70b-16bit",
		"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 "upstage/SOLAR-0-70b-16bit" \
        --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": "upstage/SOLAR-0-70b-16bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use upstage/SOLAR-0-70b-16bit with Docker Model Runner:
```
docker model run hf.co/upstage/SOLAR-0-70b-16bit
```

Call w/ LiteLLM

by krrish-litellm - opened Aug 15, 2023

Discussion

krrish-litellm

Aug 15, 2023

Hi @hunkim / @yoonniverse

What's the best way for me to deploy this model? I'd love to make a demo of this with LiteLLM - https://github.com/BerriAI/litellm.

Lite currently works with Replicate, Azure, Together.ai and HF Inference Endpoints.

I'm facing issues with HF Inference endpoints due to quota limitations, so curious if you've tried any other provider.

hunkim

upstage org Aug 15, 2023

We will soon host our model on Together.ai. We will keep you updated.

Do you know how to integrate our model with https://github.com/BerriAI/litellm? We will make it work. Let us know.

krrish-litellm

Aug 15, 2023

•

edited Aug 15, 2023

Hey @hunkim we made it easy to proxy openai with any deployment solution - should unlock any provider you choose. - https://github.com/BerriAI/litellm/issues/120

import litellm 
def translate_function(model, messages, max_tokens):
    prompt = " ".join(message["content"] for message in messages)
    max_new_tokens = max_tokens
    return {"model": model, "prompt": prompt, "max_new_tokens": max_new_tokens} 

openai.api_base = litellm.translate_api_call(custom_api_base, translate_function)

We already have a custom integration with together.ai, which supports streaming. Excited to put out a demo notebook/etc. once it's deployed.

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