[Whiper] add get_input_embeddings to WhisperForAudioClassification (#22133)

* add `get_input_embeddings` to `WhisperForAudioClassification`

* add common tests

* fix another common test

* Update tests/models/whisper/test_modeling_whisper.py

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* fix style

---------

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

src/transformers/models/whisper/modeling_whisper.py

@@ -767,6 +767,12 @@ class WhisperEncoder(WhisperPreTrainedModel):
             param.requires_grad = False
         self._requires_grad = False
 
+    def get_input_embeddings(self) -> nn.Module:
+        return self.conv1
+
+    def set_input_embeddings(self, value: nn.Module):
+        self.conv1 = value
+
     def forward(
         self,
         input_features,
@@ -1023,7 +1029,10 @@ class WhisperDecoder(WhisperPreTrainedModel):
         )
 
         # embed positions
-        positions = self.embed_positions(input_ids, past_key_values_length=past_key_values_length)
+        if input_ids is not None:
+            positions = self.embed_positions(input_ids, past_key_values_length=past_key_values_length)
+        else:
+            positions = self.embed_positions(inputs_embeds, past_key_values_length=past_key_values_length)
 
         hidden_states = inputs_embeds + positions
         hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training)
@@ -1330,6 +1339,9 @@ class WhisperForConditionalGeneration(WhisperPreTrainedModel):
     def set_output_embeddings(self, new_embeddings):
         self.proj_out = new_embeddings
 
+    def get_input_embeddings(self) -> nn.Module:
+        return self.model.get_input_embeddings()
+
     def freeze_encoder(self):
         """
         Calling this function will disable the gradient computation for the Whisper encoder so that its parameters will
@@ -1635,6 +1647,12 @@ class WhisperForAudioClassification(WhisperPreTrainedModel):
         """
         self.encoder._freeze_parameters()
 
+    def get_input_embeddings(self) -> nn.Module:
+        return self.encoder.get_input_embeddings()
+
+    def set_input_embeddings(self, value: nn.Module):
+        self.encoder.set_input_embeddings(value)
+
     @add_start_docstrings_to_model_forward(WHISPER_ENCODER_INPUTS_DOCSTRING)
     @replace_return_docstrings(output_type=SequenceClassifierOutput, config_class=_CONFIG_FOR_DOC)
     def forward(

tests/models/whisper/test_modeling_whisper.py

@@ -357,9 +357,24 @@ class WhisperModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMi
 
         return config, input_ids, None, max_length
 
-    # not implemented currently
     def test_inputs_embeds(self):
-        pass
+        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
+
+        for model_class in self.all_model_classes:
+            model = model_class(config)
+            model.to(torch_device)
+            model.eval()
+
+            inputs = copy.deepcopy(self._prepare_for_class(inputs_dict, model_class))
+
+            decoder_input_ids = inputs.pop("decoder_input_ids", None)
+            inputs.pop("decoder_attention_mask", None)
+
+            wte = model.get_input_embeddings()
+            inputs["decoder_inputs_embeds"] = wte(decoder_input_ids)
+
+            with torch.no_grad():
+                model(**inputs)[0]
 
     # training is not supported yet
     def test_training(self):
@@ -1566,9 +1581,16 @@ class WhisperEncoderModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.
 
             self.assertTrue((outputs_embeds == outputs).all())
 
-    # WhisperEncoder has no inputs_embeds and thus the `get_input_embeddings` fn is not implemented
+    # Needs to override as the encoder input embedding is a Conv1d
     def test_model_common_attributes(self):
-        pass
+        config, _ = self.model_tester.prepare_config_and_inputs_for_common()
+
+        for model_class in self.all_model_classes:
+            model = model_class(config)
+            self.assertIsInstance(model.get_input_embeddings(), (torch.nn.Conv1d))
+            model.set_input_embeddings(torch.nn.Conv1d(10, 10, 3))
+            x = model.get_output_embeddings()
+            self.assertTrue(x is None or isinstance(x, torch.nn.Conv1d))
 
     # WhisperEncoder cannot resize token embeddings since it has no tokens embeddings
     def test_resize_tokens_embeddings(self):