LFM2.5-1.2B-Base-reasoning-aggressive

REASONING-optimized | Aggressive pruning | 35% weights pruned

This model is a aggressively pruned version of LiquidAI/LFM2.5-1.2B-Base.

Performance Comparison

Category	Original	Pruned	Change
Python	0.0%	0.0%	→
Html	0.0%	0.0%	→
Trivia	20.0%	10.0%	↓ 10.0%
Math	10.0%	5.0%	↓ 5.0%
Reasoning	20.0%	5.0% ⭐	↓ 15.0%
Medical	15.0%	20.0%	↑ 5.0%
Linux	15.0%	0.0%	↓ 15.0%
Writing	5.0%	0.0%	↓ 5.0%

Average: 10.6% -> 5.0% (-5.6%)

Reasoning Retention: 25.0%

Quick Start

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("CompactAI/LFM2.5-1.2B-Base-reasoning-aggressive")
tokenizer = AutoTokenizer.from_pretrained("CompactAI/LFM2.5-1.2B-Base-reasoning-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
Specialization	Reasoning
Prune Mode	Aggressive
Weight Reduction	35% weights pruned

License

This model inherits the license from the base model.