update test_torch.py and align with common_utils.py

Author: tangleintel

tests/cpu/test_torch.py

@@ -75,8 +75,8 @@
     _compare_trilu_indices
 from common_utils import TestCase, iter_indices, TEST_NUMPY, TEST_SCIPY, TEST_MKL, \
     TEST_LIBROSA, run_tests, download_file, skipIfNoLapack, suppress_warnings, \
-    IS_WINDOWS, PY3, NO_MULTIPROCESSING_SPAWN, do_test_dtypes, do_test_empty_full, \
-    IS_SANDCASTLE, load_tests, brute_pdist, brute_cdist, slowTest, \
+    IS_WINDOWS, NO_MULTIPROCESSING_SPAWN, do_test_dtypes, do_test_empty_full, \
+    IS_SANDCASTLE, load_tests, slowTest, \
     skipCUDANonDefaultStreamIf, skipCUDAMemoryLeakCheckIf
 from multiprocessing.reduction import ForkingPickler
 from common_device_type import instantiate_device_type_tests, \
@@ -1597,9 +1597,6 @@ def _test_multinomial_invalid_probs(probs):
     @unittest.skipIf(NO_MULTIPROCESSING_SPAWN, "Disabled for environments that \
                      don't support multiprocessing with spawn start method")
     @unittest.skipIf(IS_WINDOWS, 'FIXME: CUDA OOM error on Windows')
-    @unittest.skipIf(not PY3,
-                     "spawn start method is not supported in Python 2, \
-                     but we need it for for testing failure case for CPU RNG on Windows")
     def test_multinomial_invalid_probs(self):
         test_method = _TestTorchMixin._test_multinomial_invalid_probs
         self._spawn_method(test_method, torch.Tensor([1, -1, 1]))
@@ -4002,13 +3999,10 @@ def test_serialization(self):
         buf = io.BytesIO(serialized)
         utf8_bytes = b'\xc5\xbc\xc4\x85\xc4\x85\xc3\xb3\xc5\xbc\xc4\x85\xc5\xbc'
         utf8_str = utf8_bytes.decode('utf-8')
-        if PY3:
-            loaded_utf8 = torch.load(buf, encoding='utf-8')
-            self.assertEqual(loaded_utf8, [utf8_str, torch.zeros(1, dtype=torch.float), 2])
-            buf.seek(0)
-            loaded_bytes = torch.load(buf, encoding='bytes')
-        else:
-            loaded_bytes = torch.load(buf)
+        loaded_utf8 = torch.load(buf, encoding='utf-8')
+        self.assertEqual(loaded_utf8, [utf8_str, torch.zeros(1, dtype=torch.float), 2])
+        buf.seek(0)
+        loaded_bytes = torch.load(buf, encoding='bytes')
         self.assertEqual(loaded_bytes, [utf8_bytes, torch.zeros(1, dtype=torch.float), 2])
 
     def test_serialization_filelike(self):
@@ -4292,7 +4286,6 @@ def check_map_locations(map_locations, tensor_class, intended_device):
             )
 
     @unittest.skipIf(torch.cuda.is_available(), "Testing torch.load on CPU-only machine")
-    @unittest.skipIf(not PY3, "Test tensors were serialized using python 3")
     def test_load_nonexistent_device(self):
         # Setup: create a serialized file object with a 'cuda:0' restore location
         # The following was generated by saving a torch.randn(2, device='cuda') tensor.
@@ -11025,27 +11018,6 @@ def test_nonzero_non_diff(self, device):
         nz = x.nonzero()
         self.assertFalse(nz.requires_grad)
 
-    def test_pdist_norm(self, device):
-        def test_pdist_single(shape, device, p, dtype, trans):
-            x = torch.randn(shape, dtype=dtype, device=device)
-            if trans:
-                x.transpose_(-2, -1)
-            actual = torch.pdist(x, p=p)
-            expected = brute_pdist(x, p=p)
-            self.assertEqual(expected.shape, actual.shape)
-            self.assertTrue(torch.allclose(expected, actual))
-
-        for shape in [(4, 5), (3, 2), (2, 1)]:
-            for p in [0, 1, 2, 3, 1.5, 2.5, float('inf')]:
-                for trans in [False, True]:
-                    for dtype in [torch.float32, torch.float64]:
-                        test_pdist_single(shape, device, p, dtype, trans)
-
-        # do a simplified comparison with big inputs, see:
-        # https://github.com/pytorch/pytorch/issues/15511
-        for dtype in [torch.float32, torch.float64]:
-            test_pdist_single((1000, 2), device, 2, dtype, False)
-
     def test_atan2(self, device):
         def _test_atan2_with_size(size, device):
             a = torch.rand(size=size, device=device, dtype=torch.double)