rasyosef/Phi-1_5-Instruct-v0.1

The rasyosef/Phi-1_5-Instruct-v0.1 is a 1.4 billion parameter Transformer model, fine-tuned for instruction following using supervised fine-tuning and direct preference optimization. Developed by rasyosef, it builds upon the Microsoft Phi-1.5 architecture, augmented with synthetic NLP data. This model demonstrates strong performance in common sense, language understanding, and logical reasoning, outperforming other small models on instruction following, mathematical reasoning, and general knowledge benchmarks.

Loading
Public
1.4B
BF16
2048
License: mit
Hugging Face