metadata
license: apache-2.0
tags:
- pruned
- math
- optimized
- wanda
- activation-pruning
base_model: Qwen/Qwen2.5-3B-Instruct
pipeline_tag: text-generation
Qwen2.5-3B-Instruct-math-extra-heavy
🎯 MATH-optimized | 📦 Extra Heavy pruning | ⚡ 35% weights pruned
This model is a extremely pruned version of Qwen/Qwen2.5-3B-Instruct, specialized for MATH tasks using activation-aware weight pruning (Wanda-style).
✨ Key Features
- Specialization: Optimized for Math tasks
- Pruning Method: Wanda-style (|W| × |activation|) importance scoring
- Size Reduction: 35% weights pruned
- Use Case: Maximum size reduction, best for edge deployment
📊 Performance Comparison
| Category | Original | Pruned | Change |
|---|---|---|---|
| Python | 100.0% | 100.0% | → |
| Html | 6.7% | 0.0% | ↓ 6.7% |
| Trivia | 100.0% | 66.7% | ↓ 33.3% |
| Math | 60.0% | 73.3% ⭐ | ↑ 13.3% |
| Reasoning | N/A | N/A | |
| Medical | 100.0% | 86.7% | ↓ 13.3% |
| Linux | 100.0% | 100.0% | → |
| Writing | 73.3% | 53.3% | ↓ 20.0% |
Average: 77.1% → 68.6% (-8.6%)
Math Retention: 122.2% of original performance
🚀 Quick Start
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("CompactAI/Qwen2.5-3B-Instruct-math-extra-heavy")
tokenizer = AutoTokenizer.from_pretrained("CompactAI/Qwen2.5-3B-Instruct-math-extra-heavy")
# Example usage
inputs = tokenizer("Your prompt here", return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
📋 Technical Details
| Property | Value |
|---|---|
| Base Model | Qwen/Qwen2.5-3B-Instruct |
| Specialization | Math |
| Prune Mode | Extra Heavy |
| Pruning Method | Activation-based weight pruning (Wanda) |
| Weight Reduction | 35% weights pruned |
🔗 Related Models
This model is part of the Qwen2.5-3B-Instruct pruned model collection. Other variants:
- Extra-light (minimal pruning)
- Light
- Medium-light
- Medium
- Medium-heavy
- Heavy
- Extra-heavy (maximum compression)
📜 License
This model inherits the license from the base model Qwen/Qwen2.5-3B-Instruct.
Generated by ZANNPS [Zeto Automatic Neural Network Pruning System]