check decoder_inputs_embeds is None before shifting labels (huggingfa…

…ce#19671)

src/transformers/models/bart/modeling_tf_bart.py

@@ -1352,7 +1352,7 @@ def call(
                 labels,
             )
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/blenderbot/modeling_blenderbot.py

@@ -1319,7 +1319,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/blenderbot/modeling_tf_blenderbot.py

@@ -1371,7 +1371,7 @@ def call(
                 labels,
             )
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/blenderbot_small/modeling_blenderbot_small.py

@@ -1286,7 +1286,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/blenderbot_small/modeling_tf_blenderbot_small.py

@@ -1351,7 +1351,7 @@ def call(
                 labels,
             )
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/led/modeling_led.py

@@ -2428,7 +2428,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/led/modeling_tf_led.py

@@ -2445,7 +2445,7 @@ def call(
 
         if labels is not None:
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/marian/modeling_marian.py

@@ -1432,7 +1432,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/marian/modeling_tf_marian.py

@@ -1388,7 +1388,7 @@ def call(
                 labels,
             )
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/mbart/modeling_mbart.py

@@ -1347,7 +1347,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(labels, self.config.pad_token_id)
 
         outputs = self.model(

src/transformers/models/mbart/modeling_tf_mbart.py

@@ -1387,7 +1387,7 @@ def call(
                 labels,
             )
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(labels, self.config.pad_token_id)
 
         outputs = self.model(

src/transformers/models/pegasus/modeling_pegasus.py

@@ -1393,7 +1393,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/pegasus/modeling_tf_pegasus.py

@@ -1397,7 +1397,7 @@ def call(
                 labels,
             )
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/pegasus_x/modeling_pegasus_x.py

@@ -1605,7 +1605,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/plbart/modeling_plbart.py

@@ -1314,7 +1314,7 @@ def forward(
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
         if labels is not None:
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(labels, self.config.pad_token_id)
 
         outputs = self.model(

src/transformers/models/speech_to_text/modeling_speech_to_text.py

@@ -1341,7 +1341,7 @@ def forward(
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
         if labels is not None:
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/speech_to_text/modeling_tf_speech_to_text.py

@@ -1405,7 +1405,7 @@ def call(
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
         if labels is not None:
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/whisper/modeling_tf_whisper.py

@@ -1293,7 +1293,7 @@ def call(
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
         if labels is not None:
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

src/transformers/models/whisper/modeling_whisper.py

@@ -1183,7 +1183,7 @@ def forward(
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
         if labels is not None:
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_tf_{{cookiecutter.lowercase_modelname}}.py

@@ -2938,7 +2938,7 @@ def call(
 
         if labels is not None:
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(
                     labels, self.config.pad_token_id, self.config.decoder_start_token_id
                 )

templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_{{cookiecutter.lowercase_modelname}}.py

@@ -2833,7 +2833,7 @@ def forward(
             if use_cache:
                 logger.warning("The `use_cache` argument is changed to `False` since `labels` is provided.")
             use_cache = False
-            if decoder_input_ids is None:
+            if decoder_input_ids is None and decoder_inputs_embeds is None:
                 decoder_input_ids = shift_tokens_right(labels, self.config.pad_token_id, self.config.decoder_start_token_id)
 
         outputs = self.model(