Instructions to use LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2 with Transformers:

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

pipe = pipeline("text-generation", model="LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2", device_map="auto")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2")
model = AutoModelForCausalLM.from_pretrained("LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2

SGLang

How to use LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2 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 "LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2" \
    --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": "LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2",
		"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 "LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2" \
        --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": "LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2 with Docker Model Runner:
```
docker model run hf.co/LoneStriker/CodeLlama-70b-Instruct-hf-6.0bpw-h6-exl2
```

ethically bound to not generate assembler?

by AgileTurnip - opened Jan 30, 2024

Discussion

AgileTurnip

Jan 30, 2024

codebooga has no problems with this. but codellama throws a fit:

LoneStriker

Owner Jan 30, 2024

OK, that's a bit insane if it's true. Seems like they way over-tuned a code model for "safety". Have you tried different temperature settings? A good system prompt might be a good way to guide it to just answer questions and not interject any nonsensical commentary.

AgileTurnip

Jan 30, 2024

looks like this model is uselessly broken.

https://huggingface.co/codellama/CodeLlama-70b-Instruct-hf/discussions/13

LoneStriker

Owner Jan 30, 2024

Yeah, from what I've seen, the Instruct model is hopelessly broken. We'll need to get some decent finetunes of the base model without all the guardrails and moralizing. There's a good model under the covers, but needs the community to bring it out.

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