Disable test_bert_result_with_layerwise_recompute (#18800)

### Disable test_bert_result_with_layerwise_recompute
<!-- Describe your changes. -->



### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->

orttraining/orttraining/test/python/orttraining_test_ortmodule_api.py

@@ -6396,6 +6396,9 @@ def run_step(model, x):
         del os.environ["ORTMODULE_CONV_ALGO_SEARCH"]
 
 
+@pytest.mark.skip(
+    reason="This test fail because bert forward loss is nan in updated transformers lib, disable for now."
+)
 def test_bert_result_with_layerwise_recompute():
     original_val = os.environ["ORTMODULE_MEMORY_OPT_LEVEL"] if "ORTMODULE_MEMORY_OPT_LEVEL" in os.environ else None
     # Create PyTorch model with dropout disabled.