Instructions to use openchat/openchat-3.6-8b-20240522 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use openchat/openchat-3.6-8b-20240522 with Transformers:

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

pipe = pipeline("text-generation", model="openchat/openchat-3.6-8b-20240522")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("openchat/openchat-3.6-8b-20240522")
model = AutoModelForCausalLM.from_pretrained("openchat/openchat-3.6-8b-20240522", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use openchat/openchat-3.6-8b-20240522 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "openchat/openchat-3.6-8b-20240522"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openchat/openchat-3.6-8b-20240522",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/openchat/openchat-3.6-8b-20240522

SGLang

How to use openchat/openchat-3.6-8b-20240522 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 "openchat/openchat-3.6-8b-20240522" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openchat/openchat-3.6-8b-20240522",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "openchat/openchat-3.6-8b-20240522" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openchat/openchat-3.6-8b-20240522",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use openchat/openchat-3.6-8b-20240522 with Docker Model Runner:
```
docker model run hf.co/openchat/openchat-3.6-8b-20240522
```

Chat template

by bartowski - opened May 25, 2024

Discussion

bartowski

May 25, 2024

Hey, just wanted to clarify something about the chat template

On the model card you put

GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi<|end_of_turn|>GPT4 Correct User: How are you today?<|end_of_turn|>GPT4 Correct Assistant:

but the chat_template in tokenizer_config.json compiles to:

<|start_header_id|>GPT4 Correct User<|end_header_id|>

{prompt}<|eot_id|><|start_header_id|>GPT4 Correct Assistant<|end_header_id|>

Which is correct?

count-zero

Jul 23, 2024

@bartowski I ran into the problem, that llama-server automatically selected openchat as a chat template. This caused the occurence of multiple <end_of_turn> tokens in the response.

The correct chat template for OpenChat 3.6 seems to be llama3 with the Llama 3 style EOT token: <|eot_id|>

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