Merge branch 'main' into huggingface_models_support

src/modalities/activation_checkpointing.py

@@ -11,14 +11,14 @@
 from modalities.models.gpt2.gpt2_model import GPT2Block
 
 
-def is_module_to_apply_activation_checkpointing(submodule: torch.nn.Module):
+def is_module_to_apply_activation_checkpointing(submodule: torch.nn.Module) -> bool:
     return isinstance(submodule, GPT2Block)
 
 
-def apply_activation_checkpointing_inplace(model: torch.nn.Module) -> None:
+def apply_activation_checkpointing_inplace(model: torch.nn.Module):
     assert isinstance(model, FSDP), "activation checkpointing can only be applied to FSDP wrapped models!"
-    non_reentrant_wrapper = partial(checkpoint_wrapper, checkpoint_impl=CheckpointImpl.NO_REENTRANT, debug=True)
+    non_reentrant_wrapper = partial(checkpoint_wrapper, checkpoint_impl=CheckpointImpl.NO_REENTRANT, debug=False)
 
-    return apply_activation_checkpointing(
+    apply_activation_checkpointing(
         model, checkpoint_wrapper_fn=non_reentrant_wrapper, check_fn=is_module_to_apply_activation_checkpointing
     )

src/modalities/trainer.py

@@ -58,6 +58,7 @@ def train(
         optimizer,
         loss_fun: Loss,
         callback_interval_in_batches: int,
+        # TODO: remove
         epoch_done_callback: Callable[[int], None],
         local_sample_id_to_global_sample_id: Callable[[int], int],
     ):
@@ -67,12 +68,14 @@ def train(
 
         # batch loop
         batch: DatasetBatch
+        # TODO: why do we need a barrier here?
         dist.barrier()
         forward_backward_time_recorder = TimeRecorder()
         forward_backward_time_recorder.start()
         for batch_id, batch in enumerate(train_loader):
+            # Because we might resume training, we add the starting batch id of the data loader
             local_train_batch_id = batch_id + train_loader.fast_forward_batch_id
-            # train single batch
+            # Train single batch
             batch_loss = self._train_batch(
                 batch=batch,
                 model=model,
@@ -82,7 +85,7 @@ def train(
                 data_loader=train_loader,
             )
             forward_backward_time_recorder.stop()
-            # save the batch loss
+            # Save the batch loss
             cummulated_loss[0] += batch_loss.item()
             cummulated_loss[1] += len(batch)
             batch_length_tensor = torch.tensor(len(batch)).to(torch.device(self.local_rank))