[tests] remove flax-pt equivalence and cross tests (#36283)

tests/models/big_bird/test_modeling_big_bird.py

@@ -624,15 +624,6 @@ class BigBirdModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase)
     def test_training_gradient_checkpointing_use_reentrant_false(self):
         pass
 
-    # overwrite from common in order to skip the check on `attentions`
-    def check_pt_flax_outputs(self, fx_outputs, pt_outputs, model_class, tol=1e-5, name="outputs", attributes=None):
-        # `bigbird_block_sparse_attention` in `FlaxBigBird` returns `attention_probs = None`, while in PyTorch version,
-        # an effort was done to return `attention_probs` (yet to be verified).
-        if name.startswith("outputs.attentions"):
-            return
-        else:
-            super().check_pt_flax_outputs(fx_outputs, pt_outputs, model_class, tol, name, attributes)
-
 
 @require_torch
 @slow

tests/models/big_bird/test_modeling_flax_big_bird.py

@@ -212,12 +212,3 @@ class FlaxBigBirdModelTest(FlaxModelTesterMixin, unittest.TestCase):
                 self.assertEqual(len(outputs), len(jitted_outputs))
                 for jitted_output, output in zip(jitted_outputs, outputs):
                     self.assertEqual(jitted_output.shape, output.shape)
-
-    # overwrite from common in order to skip the check on `attentions`
-    def check_pt_flax_outputs(self, fx_outputs, pt_outputs, model_class, tol=1e-5, name="outputs", attributes=None):
-        # `bigbird_block_sparse_attention` in `FlaxBigBird` returns `attention_probs = None`, while in PyTorch version,
-        # an effort was done to return `attention_probs` (yet to be verified).
-        if name.startswith("outputs.attentions"):
-            return
-        else:
-            super().check_pt_flax_outputs(fx_outputs, pt_outputs, model_class, tol, name, attributes)