add eval_packing (#1369)

docs/source/sft_trainer.mdx

@@ -271,6 +271,7 @@ trainer.train()
 ```
 
 Note that if you use a packed dataset and if you pass `max_steps` in the training arguments you will probably train your models for more than few epochs, depending on the way you have configured the packed dataset and the training protocol. Double check that you know and understand what you are doing.
+If you don't want to pack your `eval_dataset`, you can pass `eval_packing=False` to the `SFTTrainer` init method.
 
 #### Customize your prompts using packed dataset
 

tests/test_sft_trainer.py

@@ -930,3 +930,53 @@ def test_sft_trainer_tag(self):
             )
 
             assert trainer.model.model_tags == trainer._tag_names
+
+    def test_sft_trainer_eval_packing(self):
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            training_args = TrainingArguments(
+                output_dir=tmp_dir,
+                dataloader_drop_last=True,
+                evaluation_strategy="steps",
+                max_steps=4,
+                eval_steps=2,
+                save_steps=2,
+                per_device_train_batch_size=2,
+                gradient_checkpointing=True,
+            )
+
+            trainer = SFTTrainer(
+                model=self.model_id,
+                args=training_args,
+                train_dataset=self.dummy_chatml_dataset,
+                eval_dataset=self.dummy_chatml_dataset,
+                packing=True,
+                max_seq_length=32,  # make sure there is at least 1 packed sequence
+                eval_packing=False,
+            )
+
+            assert len(trainer.train_dataset["input_ids"]) == 1
+            assert len(trainer.eval_dataset["input_ids"]) != 1
+
+            trainer = SFTTrainer(
+                model=self.model_id,
+                args=training_args,
+                train_dataset=self.dummy_chatml_dataset,
+                eval_dataset=self.dummy_chatml_dataset,
+                max_seq_length=32,  # make sure there is at least 1 packed sequence
+                packing=True,
+            )
+
+            assert len(trainer.train_dataset["input_ids"]) == 1
+            assert len(trainer.eval_dataset["input_ids"]) == 1
+
+            trainer = SFTTrainer(
+                model=self.model_id,
+                args=training_args,
+                train_dataset=self.dummy_chatml_dataset,
+                eval_dataset=self.dummy_chatml_dataset,
+                max_seq_length=32,  # make sure there is at least 1 packed sequence
+                packing=False,
+            )
+
+            assert len(trainer.train_dataset["input_ids"]) != 1
+            assert len(trainer.eval_dataset["input_ids"]) != 1

trl/trainer/sft_trainer.py

@@ -116,6 +116,8 @@ class SFTTrainer(Trainer):
             Dict of Optional kwargs to pass when instantiating the model from a string
         dataset_kwargs: (`Optional[Dict]`, *optional*):
             Dict of Optional kwargs to pass when creating packed or non-packed datasets
+        eval_packing: (`Optional[bool]`, *optional*):
+            Whether to pack the eval dataset as well. Defaults to `packing` if `None` is passed.
     """
 
     _tag_names = ["trl", "sft"]
@@ -124,7 +126,7 @@ def __init__(
         self,
         model: Optional[Union[PreTrainedModel, nn.Module, str]] = None,
         args: Optional[TrainingArguments] = None,
-        data_collator: Optional[DataCollator] = None,
+        data_collator: Optional[DataCollator] = None,  # type: ignore
         train_dataset: Optional[Dataset] = None,
         eval_dataset: Optional[Union[Dataset, Dict[str, Dataset]]] = None,
         tokenizer: Optional[PreTrainedTokenizerBase] = None,
@@ -146,6 +148,7 @@ def __init__(
         neftune_noise_alpha: Optional[float] = None,
         model_init_kwargs: Optional[Dict] = None,
         dataset_kwargs: Optional[Dict] = None,
+        eval_packing: Optional[bool] = None,
     ):
         if model_init_kwargs is None:
             model_init_kwargs = {}
@@ -274,11 +277,14 @@ def make_inputs_require_grad(module, input, output):
             if eval_dataset is not None:
                 _multiple = isinstance(eval_dataset, dict)
                 _eval_datasets = eval_dataset if _multiple else {"singleton": eval_dataset}
+
+                eval_packing = packing if eval_packing is None else eval_packing
+
                 for _eval_dataset_name, _eval_dataset in _eval_datasets.items():
                     _eval_datasets[_eval_dataset_name] = self._prepare_dataset(
                         _eval_dataset,
                         tokenizer,
-                        packing,
+                        eval_packing,
                         dataset_text_field,
                         max_seq_length,
                         formatting_func,