Skip CI test to load S3 dataset on M1 due to permission denied (#882)

Summary:
Fixes #875

Pull Request resolved: #882

Reviewed By: NivekT

Differential Revision: D41028025

Pulled By: ejguan

fbshipit-source-id: 8a88da3be5b923705bbc0bfa76e4fe270803dd71

Author: ejguan

test/test_remote_io.py

@@ -13,7 +13,7 @@
 
 import torchdata
 
-from _utils._common_utils_for_test import check_hash_fn, create_temp_dir, IS_WINDOWS
+from _utils._common_utils_for_test import check_hash_fn, create_temp_dir, IS_M1, IS_WINDOWS
 from torch.utils.data import DataLoader
 
 from torchdata.datapipes.iter import (
@@ -222,6 +222,7 @@ def test_disabled_s3_io_iterdatapipe(self):
             _ = S3FileLoader(IterableWrapper(file_urls))
 
     @skipIfNoAWS
+    @unittest.skipIf(IS_M1, "PyTorch M1 CI Machine doesn't allow accessing")
     def test_s3_io_iterdatapipe(self):
         # S3FileLister: different inputs
         input_list = [