Instructions to use bartowski/Qwen2.5-Coder-7B-Instruct-exl2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bartowski/Qwen2.5-Coder-7B-Instruct-exl2 with Transformers:

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

pipe = pipeline("text-generation", model="bartowski/Qwen2.5-Coder-7B-Instruct-exl2")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("bartowski/Qwen2.5-Coder-7B-Instruct-exl2", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use bartowski/Qwen2.5-Coder-7B-Instruct-exl2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "bartowski/Qwen2.5-Coder-7B-Instruct-exl2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bartowski/Qwen2.5-Coder-7B-Instruct-exl2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/bartowski/Qwen2.5-Coder-7B-Instruct-exl2

SGLang

How to use bartowski/Qwen2.5-Coder-7B-Instruct-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 "bartowski/Qwen2.5-Coder-7B-Instruct-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": "bartowski/Qwen2.5-Coder-7B-Instruct-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 "bartowski/Qwen2.5-Coder-7B-Instruct-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": "bartowski/Qwen2.5-Coder-7B-Instruct-exl2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use bartowski/Qwen2.5-Coder-7B-Instruct-exl2 with Docker Model Runner:
```
docker model run hf.co/bartowski/Qwen2.5-Coder-7B-Instruct-exl2
```

Qwen2.5-Coder-1.5B-Instruct for speculative decoding?

by Handgun1773 - opened Sep 30, 2024

Discussion

Handgun1773

Sep 30, 2024

The title says it: iirc exl2 support speculative decoding, we have a smol qwen coder, nobody exl2ed it.
If you have the time, it would be very cool of you to do it.
Base models would also be very nice for code completion.

Handgun1773

Oct 4, 2024

Ok so I did the base to run through tabbyapi and use for completion with continue.dev:
https://huggingface.co/Handgun1773/Qwen2.5-Coder-1.5B-BASE-8.0bpw-exl2
https://huggingface.co/Handgun1773/Qwen2.5-Coder-7B-BASE-8.0bpw-exl2

Speculative decoding doesn't seem to give any noticeable boost, so I'm just using base with my inline_model_loading: true + litellm config, so I can use instruct models and code completion.

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