55 lines
2.1 KiB
Python
55 lines
2.1 KiB
Python
import os
|
|
import sys
|
|
import unittest
|
|
import torch
|
|
PROJECT_ROOT = os.path.realpath(os.path.join(os.path.dirname(__file__), "..", ".."))
|
|
MODULE_ROOT = os.path.join(PROJECT_ROOT, "modules")
|
|
sys.path.append(MODULE_ROOT)
|
|
from pytorch_wrapper.pt_to_image_transpose import PtToImageTranspose
|
|
|
|
|
|
class TestPtToImageTranspose(unittest.TestCase):
|
|
def setUp(self):
|
|
"""Set up test instance."""
|
|
self.node = PtToImageTranspose()
|
|
|
|
# Define test cases: (input_tensor, expected_shape)
|
|
self.test_cases = [
|
|
# Case 1: 3D tensor (C, H, W) -> (H, W, C)
|
|
(torch.ones(3, 64, 128), (64, 128, 3)), # (3,64,128) -> (64,128,3)
|
|
|
|
# Case 2: 4D tensor (N, C, H, W) -> (N, H, W, C)
|
|
(torch.ones(10, 3, 64, 128), (10, 64, 128, 3)), # (10,3,64,128) -> (10,64,128,3)
|
|
|
|
# Case 3: Invalid input (2D tensor)
|
|
(torch.ones(64, 128), ValueError), # 2D tensors should raise an error
|
|
|
|
# Case 4: Invalid input (5D tensor)
|
|
(torch.ones(10, 3, 64, 128, 2), ValueError), # 5D tensors should raise an error
|
|
]
|
|
|
|
def test_transpose(self):
|
|
"""Test tensor to image transposition with valid and invalid cases."""
|
|
for input_tensor, expected in self.test_cases:
|
|
with self.subTest(input_shape=input_tensor.shape):
|
|
if isinstance(expected, tuple):
|
|
# Valid cases
|
|
result, = self.node.f(input_tensor)
|
|
|
|
# Ensure return value is a tensor
|
|
self.assertTrue(isinstance(result, torch.Tensor), "Output is not a tensor")
|
|
|
|
# Ensure result has correct shape
|
|
self.assertEqual(result.shape, torch.Size(expected), f"Expected shape {expected}, got {result.shape}")
|
|
|
|
# Ensure result has the same dtype
|
|
self.assertEqual(result.dtype, input_tensor.dtype, "Dtype mismatch after transpose operation")
|
|
else:
|
|
# Invalid cases should raise an error
|
|
with self.assertRaises(expected):
|
|
self.node.f(input_tensor)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|