LLaMA-MoE: Building Mixture-of-Experts from LLaMA with Continual Pre-training

Devs

LLaMA-MoE: Building Mixture-of-Experts from LLaMA with Continual Pre-training | Read Paper on Bytez