import torch
import torch.nn as nn
import torch_npu
from torch_npu.testing.testcase import TestCase, run_tests
class TestVisionLayers(TestCase):
def test_PixelShuffle(self):
pixel_shuffle = nn.PixelShuffle(3).npu()
input1 = torch.randn(1, 9, 4, 4).npu()
output = pixel_shuffle(input1)
self.assertEqual(output is not None, True)
def test_Upsample(self):
input1 = torch.arange(1, 5, dtype=torch.float32).view(1, 1, 2, 2).npu()
m = nn.Upsample(scale_factor=2, mode='nearest').npu()
output = m(input1)
self.assertEqual(output is not None, True)
def test_UpsamplingNearest2d(self):
input1 = torch.arange(1, 5, dtype=torch.float32).view(1, 1, 2, 2).npu()
m = nn.UpsamplingNearest2d(scale_factor=2).npu()
output = m(input1)
self.assertEqual(output is not None, True)
if __name__ == "__main__":
run_tests()