README.md

---
license: apache-2.0
tags:
- pruned
- python
- optimized
base_model: Qwen/Qwen3-0.6B
---

# Qwen3-0.6B-python-light-prune

This model is a **light** pruned version of [Qwen/Qwen3-0.6B](https://huggingface.co/Qwen/Qwen3-0.6B), specialized for **PYTHON** tasks.

## Pruning Details

- **Base Model**: Qwen/Qwen3-0.6B
- **Specialization**: Python
- **Prune Mode**: Light
- **Method**: Activation-based weight pruning

## Performance Comparison

| Category | Original | Pruned |
|----------|----------|--------|
| Python | 13.3% | 20.0% |
| HTML | 53.3% | 33.3% |
| Trivia | 53.3% | 60.0% |
| Math | 40.0% | 46.7% |
| Reasoning | 33.3% | 40.0% |

![Comparison Graph](comparison_graph.png)

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("CompactAI/Qwen3-0.6B-python-light-prune-prune")
tokenizer = AutoTokenizer.from_pretrained("CompactAI/Qwen3-0.6B-python-light-prune-prune")
```

## License

This model inherits the license from the base model.