Fix model checkpoint saving issue when using PEFT, the is no check fo…

…r whether the directory already exists resulting in error when using distributed training
huggingface · Oct 26, 2024 · a14d760 · a14d760
1 parent 6a6bb03
commit a14d760
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/optimum/neuron/utils/peft_utils.py b/optimum/neuron/utils/peft_utils.py
@@ -176,7 +176,7 @@ def state_dict(self):
 
                 adapter_shards_dir_model = os.path.join(output_dir, "adapter_shards", "model")
                 if not os.path.isdir(adapter_shards_dir_model):
-                    os.makedirs(adapter_shards_dir_model)
+                    os.makedirs(adapter_shards_dir_model, exist_ok=True)
 
                 dummy_mod = DummyModule()
                 neuronx_distributed.trainer.save_checkpoint(