Falcon-H1-Tiny-R-0.6B-trivia-aggressive

TRIVIA-optimized | Aggressive pruning | 35% weights pruned

This model is a aggressively pruned version of tiiuae/Falcon-H1-Tiny-R-0.6B.

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	0.0%	0.0% ⭐	→
Math	0.0%	0.0%	→
Reasoning	0.0%	0.0%	→
Medical	0.0%	0.0%	→
Linux	0.0%	0.0%	→
Writing	0.0%	0.0%	→

Average: 0.0% -> 0.0% (+0.0%)

Quick Start

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("CompactAI/Falcon-H1-Tiny-R-0.6B-trivia-aggressive")
tokenizer = AutoTokenizer.from_pretrained("CompactAI/Falcon-H1-Tiny-R-0.6B-trivia-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	tiiuae/Falcon-H1-Tiny-R-0.6B
Specialization	Trivia
Prune Mode	Aggressive
Weight Reduction	35% weights pruned

License

This model inherits the license from the base model.