🧹 Remove max_batch_tokens, num_blocks and block_size from generation kwargs

trl/trainer/grpo_trainer.py

@@ -565,10 +565,6 @@ def __init__(
                 "repetition_penalty": self.repetition_penalty,
                 "cache_implementation": args.cache_implementation,
             }
-            if args.use_transformers_paged:
-                generation_kwargs["max_batch_tokens"] = 512
-                generation_kwargs["num_blocks"] = 1024
-                generation_kwargs["block_size"] = 128
             if args.generation_kwargs is not None:
                 generation_kwargs.update(args.generation_kwargs)
             self.generation_config = GenerationConfig(**generation_kwargs)
@@ -1305,6 +1301,7 @@ def _generate_and_score_completions(
                     all_outputs = unwrapped_model.generate_batch(
                         paged_prompt_inputs.input_ids, generation_config=self.generation_config, progress_bar=False
                     )
+                    unwrapped_model.train()  # restore training mode, as generate_batch forces eval mode
             completion_ids = [output.generated_tokens for output in all_outputs.values()]
             completion_ids = [torch.tensor(ids, device=device) for ids in completion_ids]
             completion_ids = pad(completion_ids, padding_value=self.pad_token_id, padding_side="right")

trl/trainer/online_dpo_trainer.py

@@ -571,10 +571,6 @@ def __init__(
                 generation_kwargs["min_p"] = self.min_p
             if args.generation_kwargs is not None:
                 generation_kwargs.update(args.generation_kwargs)
-            if self.use_transformers_paged:
-                generation_kwargs["max_batch_tokens"] = 512
-                generation_kwargs["num_blocks"] = 1024
-                generation_kwargs["block_size"] = 128
             # Remove None values
             generation_kwargs = {k: v for k, v in generation_kwargs.items() if v is not None}
             self.generation_config = GenerationConfig(**generation_kwargs)
@@ -1111,6 +1107,7 @@ def _generate(self, model, prompts, images=None):
                         generation_config=self.generation_config,
                         progress_bar=False,
                     )
+                    unwrapped_model.train()  # restore training mode, as generate_batch forces eval mode
             completion_ids = [output.generated_tokens for output in all_outputs.values()]
             completion_ids = [torch.tensor(ids, device=device) for ids in completion_ids]
             completion_ids = pad(completion_ids, padding_value=self.pad_token_id, padding_side="right")

trl/trainer/rloo_trainer.py

@@ -637,10 +637,6 @@ def decode(example, tokenizer):
                 "repetition_penalty": self.repetition_penalty,
                 "cache_implementation": args.cache_implementation,
             }
-            if args.use_transformers_paged:
-                generation_kwargs["max_batch_tokens"] = 512
-                generation_kwargs["num_blocks"] = 1024
-                generation_kwargs["block_size"] = 128
             if args.generation_kwargs is not None:
                 generation_kwargs.update(args.generation_kwargs)
             self.generation_config = GenerationConfig(**generation_kwargs)
@@ -1283,6 +1279,7 @@ def _generate_and_score_completions(
                     all_outputs = unwrapped_model.generate_batch(
                         paged_prompt_inputs.input_ids, generation_config=self.generation_config, progress_bar=False
                     )
+                    unwrapped_model.train()  # restore training mode, as generate_batch forces eval mode
             completion_ids = [output.generated_tokens for output in all_outputs.values()]
             completion_ids = [torch.tensor(ids, device=device) for ids in completion_ids]
             completion_ids = pad(completion_ids, padding_value=self.pad_token_id, padding_side="right")