Update tiny model information and pipeline tests (#26285)

* Update tiny model summary file

* add to pipeline tests

* revert

* fix import

* fix import

* fix

* fix

* update

* update

* update

* fix

* remove BarkModelTest

* fix

---------

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

tests/models/bark/test_modeling_bark.py

@@ -493,13 +493,6 @@ class BarkModelTester:
 
         self.is_training = is_training
 
-    def prepare_config_and_inputs(self):
-        # TODO: @Yoach: Preapre `inputs_dict`
-        inputs_dict = {}
-        config = self.get_config()
-
-        return config, inputs_dict
-
     def get_config(self):
         return BarkConfig.from_sub_model_configs(
             self.semantic_model_tester.get_config(),
@@ -522,22 +515,6 @@ class BarkModelTester:
 
         return config
 
-    def prepare_config_and_inputs_for_common(self):
-        # TODO: @Yoach
-        pass
-        # return config, inputs_dict
-
-
-# Need this class in oder to create tiny model for `bark`
-# TODO (@Yoach) Implement actual test methods
-@unittest.skip("So far all tests will fail.")
-class BarkModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase):
-    all_model_classes = (BarkModel,) if is_torch_available() else ()
-
-    def setUp(self):
-        self.model_tester = BarkModelTester(self)
-        self.config_tester = ConfigTester(self, config_class=BarkConfig, n_embd=37)
-
 
 @require_torch
 class BarkSemanticModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase):

tests/models/blip_2/test_modeling_blip_2.py

@@ -666,7 +666,11 @@ class Blip2ModelTester:
 class Blip2ModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
     all_model_classes = (Blip2ForConditionalGeneration, Blip2Model) if is_torch_available() else ()
     pipeline_model_mapping = (
-        {"feature-extraction": Blip2Model, "image-to-text": Blip2ForConditionalGeneration}
+        {
+            "feature-extraction": Blip2Model,
+            "image-to-text": Blip2ForConditionalGeneration,
+            "visual-question-answering": Blip2ForConditionalGeneration,
+        }
         if is_torch_available()
         else {}
     )

tests/models/bros/test_modeling_bros.py

@@ -22,6 +22,7 @@ from transformers.utils import is_torch_available
 
 from ...test_configuration_common import ConfigTester
 from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
+from ...test_pipeline_mixin import PipelineTesterMixin
 
 
 if is_torch_available():
@@ -272,7 +273,7 @@ class BrosModelTester:
 
 
 @require_torch
-class BrosModelTest(ModelTesterMixin, unittest.TestCase):
+class BrosModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
     test_pruning = False
     test_torchscript = False
     test_mismatched_shapes = False
@@ -288,6 +289,18 @@ class BrosModelTest(ModelTesterMixin, unittest.TestCase):
         else ()
     )
     all_generative_model_classes = () if is_torch_available() else ()
+    pipeline_model_mapping = (
+        {"feature-extraction": BrosModel, "token-classification": BrosForTokenClassification}
+        if is_torch_available()
+        else {}
+    )
+
+    # BROS requires `bbox` in the inputs which doesn't fit into the above 2 pipelines' input formats.
+    # see https://github.com/huggingface/transformers/pull/26294
+    def is_pipeline_test_to_skip(
+        self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
+    ):
+        return True
 
     def setUp(self):
         self.model_tester = BrosModelTester(self)

tests/models/idefics/test_modeling_idefics.py

@@ -260,7 +260,7 @@ class IdeficsModelTester:
 @require_torch
 class IdeficsModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
     all_model_classes = (IdeficsModel, IdeficsForVisionText2Text) if is_torch_available() else ()
-    pipeline_model_mapping = {}
+    pipeline_model_mapping = {"feature-extraction": IdeficsModel} if is_torch_available() else {}
     test_pruning = False
     test_headmasking = False
     test_torchscript = False

tests/models/pop2piano/test_modeling_pop2piano.py

@@ -37,6 +37,7 @@ from transformers.utils import is_essentia_available, is_librosa_available, is_s
 from ...generation.test_utils import GenerationTesterMixin
 from ...test_configuration_common import ConfigTester
 from ...test_modeling_common import ModelTesterMixin, ids_tensor
+from ...test_pipeline_mixin import PipelineTesterMixin
 
 
 if is_torch_available():
@@ -509,9 +510,12 @@ class Pop2PianoModelTester:
 
 
 @require_torch
-class Pop2PianoModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase):
+class Pop2PianoModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin, unittest.TestCase):
     all_model_classes = (Pop2PianoForConditionalGeneration,) if is_torch_available() else ()
     all_generative_model_classes = ()
+    pipeline_model_mapping = (
+        {"automatic-speech-recognition": Pop2PianoForConditionalGeneration} if is_torch_available() else {}
+    )
     all_parallelizable_model_classes = ()
     fx_compatible = False
     test_pruning = False

tests/models/vits/test_modeling_vits.py

@@ -156,7 +156,9 @@ class VitsModelTester:
 @require_torch
 class VitsModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
     all_model_classes = (VitsModel,) if is_torch_available() else ()
-    pipeline_model_mapping = {"text-to-audio": VitsModel} if is_torch_available() else {}
+    pipeline_model_mapping = (
+        {"feature-extraction": VitsModel, "text-to-audio": VitsModel} if is_torch_available() else {}
+    )
     is_encoder_decoder = False
     test_pruning = False
     test_headmasking = False