---
license: apache-2.0
tags:
- pruned
- math
- optimized
- wanda
base_model: LiquidAI/LFM2.5-1.2B-Base
pipeline_tag: text-generation
---

# LFM2.5-1.2B-Base-math-aggressive

> **MATH-optimized** | **Aggressive** pruning | **35% weights pruned**

This model is a **aggressively pruned** version of [LiquidAI/LFM2.5-1.2B-Base](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Base).



## Performance Comparison

| Category | Original | Pruned | Change |
|----------|----------|--------|--------|
| Python | 0.0% | 0.0% | → |
| Html | 10.0% | 0.0% | ↓ 10.0% |
| Trivia | 70.0% | 45.0% | ↓ 25.0% |
| **Math** | 35.0% | 15.0% ⭐ | ↓ 20.0% |
| Reasoning | 15.0% | 10.0% | ↓ 5.0% |
| Medical | 30.0% | 10.0% | ↓ 20.0% |
| Linux | 40.0% | 30.0% | ↓ 10.0% |
| Writing | 10.0% | 5.0% | ↓ 5.0% |

**Average**: 26.2% -> 14.4% (-11.9%)

**Math Retention**: 42.9%

![Comparison Graph](comparison_graph.png)

## Quick Start

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("CompactAI/LFM2.5-1.2B-Base-math-aggressive")
tokenizer = AutoTokenizer.from_pretrained("CompactAI/LFM2.5-1.2B-Base-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 | [LiquidAI/LFM2.5-1.2B-Base](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Base) |
| Specialization | Math |
| Prune Mode | Aggressive |
| Weight Reduction | 35% weights pruned |

## License

This model inherits the license from the base model.