metadata
license: apache-2.0
base_model: Qwen/Qwen2.5-Coder-7B-Instruct
tags:
- code-review
- python
- qwen2
- fine-tuned
datasets:
- custom
language:
- en
pipeline_tag: text-generation
Code Review Critic
A fine-tuned Qwen2.5-Coder-7B-Instruct model for Python code review.
Model Description
This model provides constructive, actionable feedback on Python code. It focuses on:
- Bug detection
- Potential issues
- Code quality improvements
Base Model: Qwen/Qwen2.5-Coder-7B-Instruct Fine-tuning Method: QLoRA (4-bit quantization + LoRA adapters) Training Data: 8,275 real GitHub PR review comments from major Python projects
Training Details
- LoRA Rank: 64
- LoRA Alpha: 64
- Learning Rate: 2e-4
- Epochs: 2
- Final Eval Loss: 0.8455
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("YOUR_USERNAME/code-review-critic")
tokenizer = AutoTokenizer.from_pretrained("YOUR_USERNAME/code-review-critic")
messages = [
{"role": "system", "content": "You are an expert code reviewer. Analyze the provided Python code and give constructive, specific feedback."},
{"role": "user", "content": "Review this Python code:\n\n```python\ndef get_user(id):\n return db.query(f'SELECT * FROM users WHERE id = {id}')\n```"}
]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))