diffusers

[train_lcm_distill_lora_sdxl.py] Fix the LR schedulers when num_train_epochs is passed in a distributed training env

#8446

Merged