Byebye test_batching_equivalence's flakiness (#35729)

* fix

* fix

* skip

* better error message

---------

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

tests/models/oneformer/test_modeling_oneformer.py

@@ -23,6 +23,7 @@ import numpy as np
 from tests.test_modeling_common import floats_tensor
 from transformers import OneFormerConfig, is_torch_available, is_vision_available
 from transformers.testing_utils import (
+    is_flaky,
     require_timm,
     require_torch,
     require_torch_accelerator,
@@ -268,6 +269,12 @@ class OneFormerModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCas
     def test_config(self):
         self.config_tester.run_common_tests()
 
+    @is_flaky(
+        description="The `attention_mask` computed with `< 0.5` in `OneFormerTransformerDecoder.forward_prediction_heads` is sensitive to input values."
+    )
+    def test_batching_equivalence(self):
+        super().test_batching_equivalence()
+
     def test_oneformer_model(self):
         config, inputs = self.model_tester.prepare_config_and_inputs_for_common()
         self.model_tester.create_and_check_oneformer_model(config, **inputs, output_hidden_states=False)