Use compute_loss in prediction_step (#9935)

src/transformers/trainer.py

@@ -1312,7 +1312,7 @@ class Trainer:
 
         return loss.detach()
 
-    def compute_loss(self, model, inputs):
+    def compute_loss(self, model, inputs, return_outputs=False):
         """
         How the loss is computed by Trainer. By default, all models return the loss in the first element.
 
@@ -1329,10 +1329,12 @@ class Trainer:
             self._past = outputs[self.args.past_index]
 
         if labels is not None:
-            return self.label_smoother(outputs, labels)
+            loss = self.label_smoother(outputs, labels)
         else:
             # We don't use .loss here since the model may return tuples instead of ModelOutput.
-            return outputs["loss"] if isinstance(outputs, dict) else outputs[0]
+            loss = outputs["loss"] if isinstance(outputs, dict) else outputs[0]
+
+        return (loss, outputs) if return_outputs else loss
 
     def is_local_process_zero(self) -> bool:
         """
@@ -1718,29 +1720,27 @@ class Trainer:
                 ignore_keys = []
 
         with torch.no_grad():
-            if self.use_amp:
-                with autocast():
-                    outputs = model(**inputs)
-            else:
-                outputs = model(**inputs)
             if has_labels:
-                if self.label_smoother is not None and "labels" in inputs:
+                loss, outputs = self.compute_loss(model, inputs, return_outputs=True)
-                    loss = self.label_smoother(outputs, inputs["labels"]).mean().detach()
+                loss = loss.mean().detach()
-                else:
-                    loss = (outputs["loss"] if isinstance(outputs, dict) else outputs[0]).mean().detach()
                 if isinstance(outputs, dict):
                     logits = tuple(v for k, v in outputs.items() if k not in ignore_keys + ["loss"])
                 else:
                     logits = outputs[1:]
             else:
                 loss = None
+                if self.use_amp:
+                    with autocast():
+                        outputs = model(**inputs)
+                else:
+                    outputs = model(**inputs)
                 if isinstance(outputs, dict):
                     logits = tuple(v for k, v in outputs.items() if k not in ignore_keys)
                 else:
                     logits = outputs
                 # TODO: this needs to be fixed and made cleaner later.
                 if self.args.past_index >= 0:
-                self._past = outputs[self.args.past_index if has_labels else self.args.past_index - 1]
+                    self._past = outputs[self.args.past_index - 1]
 
         if prediction_loss_only:
             return (loss, None, None)