Fix the inconsistency of loss calculation between PT/TF XLNetLMHeadModel (#15298)

* Fix the inconsistency of loss calculation between PT/TF XLNetLMHeadModel

* overwrite test_loss_computation

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

src/transformers/models/xlnet/modeling_tf_xlnet.py

@@ -1390,10 +1390,7 @@ class TFXLNetLMHeadModel(TFXLNetPreTrainedModel, TFCausalLanguageModelingLoss):
 
         loss = None
         if inputs["labels"] is not None:
-            # shift labels to the left and cut last logit token
+            loss = self.hf_compute_loss(inputs["labels"], logits)
-            logits = logits[:, :-1]
-            labels = inputs["labels"][:, 1:]
-            loss = self.hf_compute_loss(labels, logits)
 
         if not inputs["return_dict"]:
             output = (logits,) + transformer_outputs[1:]

tests/test_modeling_tf_xlnet.py

@@ -14,6 +14,7 @@
 # limitations under the License.
 
 
+import inspect
 import random
 import unittest
 
@@ -391,6 +392,69 @@ class TFXLNetModelTest(TFModelTesterMixin, unittest.TestCase):
             model = TFXLNetModel.from_pretrained(model_name)
             self.assertIsNotNone(model)
 
+    # overwrite since `TFXLNetLMHeadModel` doesn't cut logits/labels
+    def test_loss_computation(self):
+        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
+        for model_class in self.all_model_classes:
+            model = model_class(config)
+            if getattr(model, "hf_compute_loss", None):
+                # The number of elements in the loss should be the same as the number of elements in the label
+                prepared_for_class = self._prepare_for_class(inputs_dict.copy(), model_class, return_labels=True)
+                added_label = prepared_for_class[
+                    sorted(list(prepared_for_class.keys() - inputs_dict.keys()), reverse=True)[0]
+                ]
+                loss_size = tf.size(added_label)
+
+                # `TFXLNetLMHeadModel` doesn't cut logits/labels
+                # if model.__class__ in get_values(TF_MODEL_FOR_CAUSAL_LM_MAPPING):
+                #     # if loss is causal lm loss, labels are shift, so that one label per batch
+                #     # is cut
+                #     loss_size = loss_size - self.model_tester.batch_size
+
+                # Test that model correctly compute the loss with kwargs
+                prepared_for_class = self._prepare_for_class(inputs_dict.copy(), model_class, return_labels=True)
+                input_name = "input_ids" if "input_ids" in prepared_for_class else "pixel_values"
+                input_ids = prepared_for_class.pop(input_name)
+
+                loss = model(input_ids, **prepared_for_class)[0]
+                self.assertEqual(loss.shape, [loss_size])
+
+                # Test that model correctly compute the loss with a dict
+                prepared_for_class = self._prepare_for_class(inputs_dict.copy(), model_class, return_labels=True)
+                loss = model(prepared_for_class)[0]
+                self.assertEqual(loss.shape, [loss_size])
+
+                # Test that model correctly compute the loss with a tuple
+                prepared_for_class = self._prepare_for_class(inputs_dict.copy(), model_class, return_labels=True)
+
+                # Get keys that were added with the _prepare_for_class function
+                label_keys = prepared_for_class.keys() - inputs_dict.keys()
+                signature = inspect.signature(model.call).parameters
+                signature_names = list(signature.keys())
+
+                # Create a dictionary holding the location of the tensors in the tuple
+                tuple_index_mapping = {0: input_name}
+                for label_key in label_keys:
+                    label_key_index = signature_names.index(label_key)
+                    tuple_index_mapping[label_key_index] = label_key
+                sorted_tuple_index_mapping = sorted(tuple_index_mapping.items())
+                # Initialize a list with their default values, update the values and convert to a tuple
+                list_input = []
+
+                for name in signature_names:
+                    if name != "kwargs":
+                        list_input.append(signature[name].default)
+
+                for index, value in sorted_tuple_index_mapping:
+                    list_input[index] = prepared_for_class[value]
+
+                tuple_input = tuple(list_input)
+
+                # Send to model
+                loss = model(tuple_input[:-1])[0]
+
+                self.assertEqual(loss.shape, [loss_size])
+
 
 @require_tf
 class TFXLNetModelLanguageGenerationTest(unittest.TestCase):