mjbooo

### Prerequisites * [X] Put an X between the brackets on this line if you have done all of the following: * Checked the online documentation: https://mimic.mit.edu/ * Checked that...

https://github.com/huggingface/trl/blob/0ee349dcd43b0f4b3169449f16751c38ac4a609f/trl/trainer/dpo_trainer.py#L758C1-L771C76 ``` def compute_reference_log_probs(self, padded_batch: Dict) -> Dict: """Computes log probabilities of the reference model for a single padded batch of a DPO specific dataset.""" compte_ref_context_manager = torch.cuda.amp.autocast if self._peft_has_been_casted_to_bf16...