alibaba / TinyNeuralNetwork

TinyNeuralNetwork is an efficient and easy-to-use deep learning model compression framework.
MIT License
747 stars 115 forks source link

[CI] several tests for modifier failed #260

Open peterjc123 opened 1 year ago

peterjc123 commented 1 year ago
FAILED tests/modifier_speed_test.py::ModifierForwardTester::test_shufflenet - RuntimeError: Given groups=1, weight of size [12, 11, 1, 1], expected input[1, 12, 28, 28] to have 11 channels, but got 12 channels instead
FAILED tests/modifier_torchvision_test.py::TestModel::test_torchvision_model_inception_v3 - AssertionError
FAILED tests/modifier_torchvision_test.py::TestModel::test_torchvision_model_shufflenet_v2_x0_5 - RuntimeError: Given groups=1, weight of size [12, 11, 1, 1], expected input[1, 12, 28, 28] to have 11 channels, but got 12 channels instead
FAILED tests/modifier_torchvision_test.py::TestModel::test_torchvision_model_shufflenet_v2_x1_0 - RuntimeError: Given groups=1, weight of size [29, 28, 1, 1], expected input[1, 29, 28, 28] to have 28 channels, but got 29 channels instead
FAILED tests/modifier_torchvision_test.py::TestModel::test_torchvision_model_shufflenet_v2_x1_5 - RuntimeError: Given groups=1, weight of size [44, 43, 1, 1], expected input[1, 44, 28, 28] to have 43 channels, but got 44 channels instead
FAILED tests/modifier_torchvision_test.py::TestModel::test_torchvision_model_shufflenet_v2_x2_0 - RuntimeError: Given groups=1, weight of size [61, 60, 1, 1], expected input[1, 61, 28, 28] to have 60 channels, but got 61 channels instead
peterjc123 commented 5 months ago

These tests are skipped in https://github.com/alibaba/TinyNeuralNetwork/pull/304.