CI: avoid human error, automatically infer generative models (#33212)

* tmp commit

* move tests to the right class

* remove ALL all_generative_model_classes = ...

* skip tf roberta

* skip InstructBlipForConditionalGenerationDecoderOnlyTest

* videollava

* reduce diff

* reduce diff

* remove  on vlms

* fix a few more

* manual rebase bits

* more manual rebase

* remove all manual generative model class test entries

* fix up to ernie

* a few more removals

* handle remaining cases

* recurrent gemma

* it's better here

* make fixup

* tf idefics is broken

* tf bert + generate is broken

* don't touch tf :()

* don't touch tf :(

* make fixup

* better comments for test skips

* revert tf changes

* remove empty line removal

* one more

* missing one

tests/models/bark/test_modeling_bark.py

@@ -22,6 +22,7 @@ import unittest
 import pytest
 
 from transformers import (
+    BarkCausalModel,
     BarkCoarseConfig,
     BarkConfig,
     BarkFineConfig,
@@ -53,7 +54,6 @@ if is_torch_available():
     import torch
 
     from transformers import (
-        BarkCausalModel,
         BarkCoarseModel,
         BarkFineModel,
         BarkModel,
@@ -527,6 +527,8 @@ class BarkModelTester:
 @require_torch
 class BarkSemanticModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase):
     all_model_classes = (BarkSemanticModel,) if is_torch_available() else ()
+    # `BarkSemanticModel` inherits from `BarkCausalModel`, but requires an advanced generation config.
+    # `BarkCausalModel` does not, so we run generation tests there.
     all_generative_model_classes = (BarkCausalModel,) if is_torch_available() else ()
 
     is_encoder_decoder = False
@@ -614,8 +616,9 @@ class BarkSemanticModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.Te
 
 @require_torch
 class BarkCoarseModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase):
-    # Same tester as BarkSemanticModelTest, except for model_class and config_class
     all_model_classes = (BarkCoarseModel,) if is_torch_available() else ()
+    # `BarkCoarseModel` inherits from `BarkCausalModel`, but requires an advanced generation config.
+    # `BarkCausalModel` does not, so we run generation tests there.
     all_generative_model_classes = (BarkCausalModel,) if is_torch_available() else ()
 
     is_encoder_decoder = False