EFFICIENT GRPO AT SCALE

Summary from the TLDR AI newsletter

Liger is an optimized trainer for Group Relative Policy Optimization (GRPO) that cuts memory usage by 40% and adds FSDP and PEFT support, making reinforcement learning fine-tuning more efficient and scalable.