update

modelscope · Oct 31, 2024 · d8c1d8d · d8c1d8d
1 parent 280edc3
commit d8c1d8d
Showing 1 changed file with 3 additions and 0 deletions.
diff --git a/swift/trainers/trainers.py b/swift/trainers/trainers.py
@@ -159,6 +159,9 @@ def compute_loss(self, model, inputs, return_outputs=None, **kwargs):
 
         loss_kwargs['labels'] = labels
         outputs = model(**inputs)
+        # fix transformers>=4.46
+        if outputs.loss is not None and 'num_items_in_batch' in kwargs:
+            outputs.loss = outputs.loss * (inputs['labels'] != -100).sum() / kwargs['num_items_in_batch']
         if loss_name is not None:
             loss_func = get_loss_func(loss_name)
             outputs['loss'] = loss_func(outputs, **loss_kwargs)