fix DPO + GC issues (huggingface#927)

trl/trainer/dpo_trainer.py

@@ -138,7 +138,30 @@ def __init__(
         elif is_peft_available() and peft_config is not None:
             if getattr(model, "is_loaded_in_8bit", False) or getattr(model, "is_loaded_in_4bit", False):
                 model = prepare_model_for_kbit_training(model, use_gradient_checkpointing=args.gradient_checkpointing)
+            elif getattr(args, "gradient_checkpointing", False):
+                # For backward compatibility with older versions of transformers
+                if hasattr(model, "enable_input_require_grads"):
+                    model.enable_input_require_grads()
+                else:
+
+                    def make_inputs_require_grad(module, input, output):
+                        output.requires_grad_(True)
+
+                    model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)
             model = get_peft_model(model, peft_config)
+        # For models that use gradient_checkpoiting, we need to attach a hook that enables input
+        # to explicitly have `requires_grad=True`, otherwise training will either silently
+        # fail or completely fail.
+        elif getattr(args, "gradient_checkpointing", False):
+            # For backward compatibility with older versions of transformers
+            if hasattr(model, "enable_input_require_grads"):
+                model.enable_input_require_grads()
+            else:
+
+                def make_inputs_require_grad(module, input, output):
+                    output.requires_grad_(True)
+
+                model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)
 
         if generate_during_eval and not is_wandb_available():
             raise ValueError(