Use labels to remove deprecation warnings (#4807)

tests/test_modeling_albert.py

@@ -162,7 +162,7 @@ class AlbertModelTest(ModelTesterMixin, unittest.TestCase):
                 input_ids,
                 attention_mask=input_mask,
                 token_type_ids=token_type_ids,
-                masked_lm_labels=token_labels,
+                labels=token_labels,
                 sentence_order_label=sequence_labels,
             )
             result = {
@@ -183,7 +183,7 @@ class AlbertModelTest(ModelTesterMixin, unittest.TestCase):
             model.to(torch_device)
             model.eval()
             loss, prediction_scores = model(
-                input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, masked_lm_labels=token_labels
+                input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels
             )
             result = {
                 "loss": loss,

tests/test_modeling_bart.py

@@ -296,7 +296,7 @@ class BartTranslationTests(unittest.TestCase):
         lm_model = BartForConditionalGeneration(config).to(torch_device)
         context = torch.Tensor([[71, 82, 18, 33, 46, 91, 2], [68, 34, 26, 58, 30, 2, 1]]).long().to(torch_device)
         summary = torch.Tensor([[82, 71, 82, 18, 2], [58, 68, 2, 1, 1]]).long().to(torch_device)
-        loss, logits, enc_features = lm_model(input_ids=context, decoder_input_ids=summary, lm_labels=summary)
+        loss, logits, enc_features = lm_model(input_ids=context, decoder_input_ids=summary, labels=summary)
         expected_shape = (*summary.shape, config.vocab_size)
         self.assertEqual(logits.shape, expected_shape)
 
@@ -361,7 +361,7 @@ class BartHeadTests(unittest.TestCase):
         lm_labels = ids_tensor([batch_size, input_ids.shape[1]], self.vocab_size).to(torch_device)
         lm_model = BartForConditionalGeneration(config)
         lm_model.to(torch_device)
-        loss, logits, enc_features = lm_model(input_ids=input_ids, lm_labels=lm_labels)
+        loss, logits, enc_features = lm_model(input_ids=input_ids, labels=lm_labels)
         expected_shape = (batch_size, input_ids.shape[1], config.vocab_size)
         self.assertEqual(logits.shape, expected_shape)
         self.assertIsInstance(loss.item(), float)
@@ -381,7 +381,7 @@ class BartHeadTests(unittest.TestCase):
         lm_model = BartForConditionalGeneration(config).to(torch_device)
         context = torch.Tensor([[71, 82, 18, 33, 46, 91, 2], [68, 34, 26, 58, 30, 2, 1]]).long().to(torch_device)
         summary = torch.Tensor([[82, 71, 82, 18, 2], [58, 68, 2, 1, 1]]).long().to(torch_device)
-        loss, logits, enc_features = lm_model(input_ids=context, decoder_input_ids=summary, lm_labels=summary)
+        loss, logits, enc_features = lm_model(input_ids=context, decoder_input_ids=summary, labels=summary)
         expected_shape = (*summary.shape, config.vocab_size)
         self.assertEqual(logits.shape, expected_shape)
 

tests/test_modeling_bert.py

@@ -218,7 +218,7 @@ class BertModelTester:
         model.to(torch_device)
         model.eval()
         loss, prediction_scores = model(
-            input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, masked_lm_labels=token_labels
+            input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels
         )
         result = {
             "loss": loss,
@@ -248,7 +248,7 @@ class BertModelTester:
             input_ids,
             attention_mask=input_mask,
             token_type_ids=token_type_ids,
-            masked_lm_labels=token_labels,
+            labels=token_labels,
             encoder_hidden_states=encoder_hidden_states,
             encoder_attention_mask=encoder_attention_mask,
         )
@@ -256,7 +256,7 @@ class BertModelTester:
             input_ids,
             attention_mask=input_mask,
             token_type_ids=token_type_ids,
-            masked_lm_labels=token_labels,
+            labels=token_labels,
             encoder_hidden_states=encoder_hidden_states,
         )
         result = {
@@ -294,7 +294,7 @@ class BertModelTester:
             input_ids,
             attention_mask=input_mask,
             token_type_ids=token_type_ids,
-            masked_lm_labels=token_labels,
+            labels=token_labels,
             next_sentence_label=sequence_labels,
         )
         result = {

tests/test_modeling_distilbert.py

@@ -151,7 +151,7 @@ class DistilBertModelTest(ModelTesterMixin, unittest.TestCase):
             model = DistilBertForMaskedLM(config=config)
             model.to(torch_device)
             model.eval()
-            loss, prediction_scores = model(input_ids, attention_mask=input_mask, masked_lm_labels=token_labels)
+            loss, prediction_scores = model(input_ids, attention_mask=input_mask, labels=token_labels)
             result = {
                 "loss": loss,
                 "prediction_scores": prediction_scores,

tests/test_modeling_electra.py

@@ -180,7 +180,7 @@ class ElectraModelTest(ModelTesterMixin, unittest.TestCase):
             model.to(torch_device)
             model.eval()
             loss, prediction_scores = model(
-                input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, masked_lm_labels=token_labels
+                input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels
             )
             result = {
                 "loss": loss,

tests/test_modeling_gpt2.py

@@ -268,7 +268,7 @@ class GPT2ModelTest(ModelTesterMixin, unittest.TestCase):
                 "mc_token_ids": mc_token_ids,
                 "attention_mask": multiple_choice_input_mask,
                 "token_type_ids": multiple_choice_token_type_ids,
-                "lm_labels": multiple_choice_inputs_ids,
+                "labels": multiple_choice_inputs_ids,
             }
 
             loss, lm_logits, mc_logits, _ = model(**inputs)

tests/test_modeling_longformer.py

@@ -164,7 +164,7 @@ class LongformerModelTester(object):
         model.to(torch_device)
         model.eval()
         loss, prediction_scores = model(
-            input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, masked_lm_labels=token_labels
+            input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels
         )
         result = {
             "loss": loss,
@@ -361,7 +361,7 @@ class LongformerModelIntegrationTest(unittest.TestCase):
             [[0] + [20920, 232, 328, 1437] * 1000 + [2]], dtype=torch.long, device=torch_device
         )  # long input
 
-        loss, prediction_scores = model(input_ids, masked_lm_labels=input_ids)
+        loss, prediction_scores = model(input_ids, labels=input_ids)
 
         expected_loss = torch.tensor(0.0620, device=torch_device)
         expected_prediction_scores_sum = torch.tensor(-6.1599e08, device=torch_device)

tests/test_modeling_openai.py

@@ -169,7 +169,7 @@ class OpenAIGPTModelTest(ModelTesterMixin, unittest.TestCase):
             model.to(torch_device)
             model.eval()
 
-            loss, lm_logits, mc_logits = model(input_ids, token_type_ids=token_type_ids, lm_labels=input_ids)
+            loss, lm_logits, mc_logits = model(input_ids, token_type_ids=token_type_ids, labels=input_ids)
 
             result = {"loss": loss, "lm_logits": lm_logits}
 

tests/test_modeling_roberta.py

@@ -155,7 +155,7 @@ class RobertaModelTest(ModelTesterMixin, unittest.TestCase):
             model.to(torch_device)
             model.eval()
             loss, prediction_scores = model(
-                input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, masked_lm_labels=token_labels
+                input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels
             )
             result = {
                 "loss": loss,

tests/test_modeling_t5.py

@@ -206,7 +206,7 @@ class T5ModelTest(ModelTesterMixin, unittest.TestCase):
                 input_ids=input_ids,
                 decoder_input_ids=decoder_input_ids,
                 decoder_attention_mask=decoder_attention_mask,
-                lm_labels=lm_labels,
+                labels=lm_labels,
             )
             loss, prediction_scores, _, _ = outputs
             self.parent.assertEqual(len(outputs), 4)