rhysjones/phi-2-orange

rhysjones/phi-2-orange is a 3 billion parameter language model, a two-step fine-tune of Microsoft's Phi-2 architecture, developed by rhysjones. This model is enhanced through broad training data and DPO fine-tuning, demonstrating competitive performance in general language understanding and reasoning tasks. It is designed for general-purpose applications requiring a compact yet capable language model with a 2048-token context length.

Cold
Public
3B
BF16
2048
License: mit
Hugging Face