|
|
--- |
|
|
license: apache-2.0 |
|
|
tags: |
|
|
- pruned |
|
|
- python |
|
|
- optimized |
|
|
- wanda |
|
|
base_model: LiquidAI/LFM2.5-1.2B-Thinking |
|
|
pipeline_tag: text-generation |
|
|
--- |
|
|
|
|
|
# LFM2.5-1.2B-Thinking-python-safe |
|
|
|
|
|
> **PYTHON-optimized** | **Safe** pruning | **30% weights pruned** |
|
|
|
|
|
This model is a **conservatively pruned** version of [LiquidAI/LFM2.5-1.2B-Thinking](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Thinking). |
|
|
|
|
|
|
|
|
|
|
|
> **Pruning Alert:** The benchmarks show virtually NO quality drop! This isn't a bug -- it is a feature. The Wanda pruning algorithm is so effective at identifying unimportant weights that it can remove a large percentage of parameters without affecting performance. Think of it like pruning dead leaves from a tree -- the tree does not miss them because they were not doing anything anyway! |
|
|
|
|
|
|
|
|
|
|
|
## Performance Comparison |
|
|
|
|
|
| Category | Original | Pruned | Change | |
|
|
|----------|----------|--------|--------| |
|
|
| **Python** | 0.0% | 0.0% ⭐ | → | |
|
|
| Html | 0.0% | 0.0% | → | |
|
|
| Trivia | 75.0% | 90.0% | ↑ 15.0% | |
|
|
| Math | 95.0% | 85.0% | ↓ 10.0% | |
|
|
| Reasoning | 85.0% | 80.0% | ↓ 5.0% | |
|
|
| Medical | 80.0% | 80.0% | → | |
|
|
| Linux | 65.0% | 55.0% | ↓ 10.0% | |
|
|
| Writing | 45.0% | 50.0% | ↑ 5.0% | |
|
|
|
|
|
**Average**: 55.6% -> 55.0% (-0.6%) |
|
|
|
|
|
|
|
|
|
|
|
 |
|
|
|
|
|
## Quick Start |
|
|
|
|
|
```python |
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
|
|
|
model = AutoModelForCausalLM.from_pretrained("CompactAI/LFM2.5-1.2B-Thinking-python-safe") |
|
|
tokenizer = AutoTokenizer.from_pretrained("CompactAI/LFM2.5-1.2B-Thinking-python-safe") |
|
|
|
|
|
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-Thinking](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Thinking) | |
|
|
| Specialization | Python | |
|
|
| Prune Mode | Safe | |
|
|
| Weight Reduction | 30% weights pruned | |
|
|
|
|
|
## License |
|
|
|
|
|
This model inherits the license from the base model. |
|
|
|
