Closed ndem0 closed 1 month ago
Describe the bug
_______________________________ test_forward[V2] _______________________________ V = tensor([[-2.2569e-01, -1.5197e-01, -1.2952e+00, 8.2365e-01, -1.3104e-02], [ 5.5923e-01, -1.9962e+00, 1.2643e...01, -1.1762e+00, 2.0898e-01, 6.8500e-03], [-4.8608e-01, -1.4008e-01, 1.9982e-01, 1.1502e+00, 1.2391e-02]]) @pytest.mark.parametrize("V", list_matrices) def test_forward(V): orth = OrthogonalBlock() orth_row = OrthogonalBlock(0) V_orth = orth(V) V_orth_row = orth_row(V.T) > assert torch.allclose(V_orth.T @ V_orth, torch.eye(V.shape[1]), atol=1e-6) E assert False E + where False = <built-in method allclose of type object at 0x7fb32ec67500>((tensor([[-0.1476, 0.3658, -0.4417, 0.7404, -0.3179],\n [-0.0706, -0.9273, -0.2317, 0.2778, -0.0651],\n ...0.0844],\n [ 0.5195, 0.0561, -0.4283, 0.1318, 0.[72](https://github.com/mathLab/PINA/actions/runs/10421892539/job/29240913616?pr=327#step:5:73)54],\n [-0.6357, 0.0179, 0.3517, 0.3323, 0.6012]]) @ tensor([[-0.1476, -0.0706, -0.5470, 0.5195, -0.6357],\n [ 0.3658, -0.92[73](https://github.com/mathLab/PINA/actions/runs/10421892539/job/29240913616?pr=327#step:5:74), 0.0534, 0.0561, 0.0179],\n ...0.3517],\n [ 0.7404, 0.2778, -0.4968, 0.1318, 0.3323],\n [-0.3179, -0.0651, 0.0844, 0.7254, 0.6012]])), tensor([[1., 0., 0., 0., 0.],\n [0., 1., 0., 0., 0.],\n [0., 0., 1., 0., 0.],\n [0., 0., 0., 1., 0.],\n [0., 0., 0., 0., 1.]]), atol=1e-06) E + where <built-in method allclose of type object at 0x7fb32ec67500> = torch.allclose E + and tensor([[-0.1476, 0.3658, -0.4417, 0.[74](https://github.com/mathLab/PINA/actions/runs/10421892539/job/29240913616?pr=327#step:5:75)04, -0.3179],\n [-0.0706, -0.9273, -0.2317, 0.2778, -0.0651],\n ...0.0844],\n [ 0.5195, 0.0561, -0.4283, 0.1318, 0.7254],\n [-0.6357, 0.0179, 0.3517, 0.3323, 0.6012]]) = tensor([[-0.1476, -0.0706, -0.5470, 0.5195, -0.6357],\n [ 0.3658, -0.9273, 0.0534, 0.0561, 0.0179],\n ...0.3517],\n [ 0.7404, 0.2778, -0.4968, 0.1318, 0.3323],\n [-0.3179, -0.0651, 0.0844, 0.7254, 0.6012]]).T E + and tensor([[1., 0., 0., 0., 0.],\n [0., 1., 0., 0., 0.],\n [0., 0., 1., 0., 0.],\n [0., 0., 0., 1., 0.],\n [0., 0., 0., 0., 1.]]) = <built-in method eye of type object at 0x7fb32ec6[75](https://github.com/mathLab/PINA/actions/runs/10421892539/job/29240913616?pr=327#step:5:76)00>(5) E + where <built-in method eye of type object at 0x7fb32ec67500> = torch.eye tests/test_layers/test_orthogonal.py:24: AssertionError
To Reproduce PR #327
Additional context Only on Win3.11 and Ubuntu3.9
Fixed in #332
Describe the bug
To Reproduce PR #327
Additional context Only on Win3.11 and Ubuntu3.9