skip OOM

egs/speech_llm/ASR_LLM/zipformer_llm_zh/train.py

@@ -32,6 +32,7 @@ torchrun --nproc_per_node 8 ./zipformer_llm_zh/train.py \
 """
 
 import argparse
+import gc
 import logging
 import os
 import warnings
@@ -625,6 +626,12 @@ def train_one_epoch(
         The rank of the node in DDP training. If no DDP is used, it should
         be set to 0.
     """
+
+    def free_gpu_cache():
+        gc.collect()
+        if torch.cuda.is_available():
+            torch.cuda.empty_cache()
+
     model.train()
     model.encoder.eval()
     if not params.unfreeze_llm:
@@ -688,9 +695,6 @@ def train_one_epoch(
                 batch=batch,
                 is_training=True,
             )
-            # summary stats
-            tot_loss = (tot_loss * (1 - 1 / params.reset_interval)) + loss_info
-
             # NOTE: We use reduction==sum and loss is computed over utterances
             # in the batch and there is no normalization to it so far.
 
@@ -700,9 +704,26 @@ def train_one_epoch(
             model.backward(loss)
             model.step()
 
-        except:  # noqa
+            # summary stats
+            tot_loss = (tot_loss * (1 - 1 / params.reset_interval)) + loss_info
+
+        except Exception as e:
+            logging.warning(f"Caught exception: {e}")
+            if (
+                "CUDA" not in str(e)
+                and "cuDNN error" not in str(e)
+                and "NCCL error" not in str(e)
+            ):
                 display_and_save_batch(batch, params=params)
-            raise
+                raise e
+
+            try:
+                loss = None
+                loss_info = None
+            except:
+                pass
+
+            free_gpu_cache()
 
         if batch_idx % params.log_interval == 0:
             try: