Instructions to use beowolx/MistralHermes-CodePro-7B-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use beowolx/MistralHermes-CodePro-7B-v1 with Transformers:

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

pipe = pipeline("text-generation", model="beowolx/MistralHermes-CodePro-7B-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("beowolx/MistralHermes-CodePro-7B-v1")
model = AutoModelForCausalLM.from_pretrained("beowolx/MistralHermes-CodePro-7B-v1")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use beowolx/MistralHermes-CodePro-7B-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "beowolx/MistralHermes-CodePro-7B-v1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "beowolx/MistralHermes-CodePro-7B-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/beowolx/MistralHermes-CodePro-7B-v1

SGLang

How to use beowolx/MistralHermes-CodePro-7B-v1 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 "beowolx/MistralHermes-CodePro-7B-v1" \
    --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": "beowolx/MistralHermes-CodePro-7B-v1",
		"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 "beowolx/MistralHermes-CodePro-7B-v1" \
        --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": "beowolx/MistralHermes-CodePro-7B-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use beowolx/MistralHermes-CodePro-7B-v1 with Docker Model Runner:
```
docker model run hf.co/beowolx/MistralHermes-CodePro-7B-v1
```

Any eval results?

by geepytee - opened Dec 13, 2023

Discussion

geepytee

Dec 13, 2023

Would be great to see humaneval results for this fine tune.

geepytee

Dec 14, 2023

Just tried running this fine tune in a Space, was hoping to put it through humaneval but the results don't really make sense even as I was testing basic tasks. Even a simple "write a python function that prints hello world" gets the LLM rambling. Perhaps the person who fine tune it can provide some guidance?

beowolx

Owner Dec 14, 2023

it uses the same prompt template the OpenHermes 2.5, which is ChatML
tried with that: https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B#prompt-format

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