Qwen3-0.6B-math-aggressive

🎯 MATH-optimized | 📦 Aggressive pruning | ⚡ 35% weights pruned

This model is a aggressively pruned version of Qwen/Qwen3-0.6B.

Performance Comparison

Category	Original	Pruned	Change
Python	0.0%	0.0%	→
Html	0.0%	0.0%	→
Trivia	28.6%	7.1%	↓ 21.4%
Math	33.3%	80.0% ⭐	↑ 46.7%
Reasoning	20.0%	20.0%	→
Medical	66.7%	16.7%	↓ 50.0%
Linux	40.0%	40.0%	→
Writing	16.7%	0.0%	↓ 16.7%

Average: 25.7% → 20.5% (-5.2%)

Math Retention: 240.0%

Quick Start

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("CompactAI/Qwen3-0.6B-math-aggressive")
tokenizer = AutoTokenizer.from_pretrained("CompactAI/Qwen3-0.6B-math-aggressive")

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/Qwen3-0.6B
Specialization	Math
Prune Mode	Aggressive
Weight Reduction	35% weights pruned

License

This model inherits the license from the base model.