fix

Author: NINGBENZHE

transfer_queue/dataloader/streaming_dataset.py

@@ -202,7 +202,7 @@ def __iter__(self) -> Iterator[tuple[TensorDict, BatchMeta]]:
                 if self.batch_index <= len(self.buffer) - 1:
                     current_data = self.buffer[self.batch_index]
                     self.batch_index += 1
-                    logger.info(f"StreamDataloader current batch index is {self.batch_index}/{len(self.buffer)}")
+                    logger.debug(f"StreamDataloader current batch index is {self.batch_index}/{len(self.buffer)}")
                     yield from self.process_batch_fn(*current_data, micro_batch_size=self.micro_batch_size)
 
                 else:

transfer_queue/sampler/seqlen_balanced_sampler.py

@@ -304,7 +304,7 @@ def __repr__(self) -> str:
 
 def get_seqlen_balanced_partitions(seqlen_list: list[int], k_partitions: int, equal_size: bool):
     """get order of seq lengths to make partitions balanced, this is
-        used in balacing sum of seqlength across dp ranks and microbatches
+        used in balancing sum of seqlength across dp ranks and microbatches
     Parameters:
         seqlen_list (List[int]):
             seq lengths of each items