Skip to content

test_linalg.py not run in CI #45271

@antocuni

Description

@antocuni

🐛 Bug

I think that test/test_linalg.py is never run by any CI build.
It is not listed inside test.run_tests.TESTS.

To Reproduce

I noticed because it fails on my machine:

_____ TestLinalgCPU.test_det_cpu_float64 _____

self = <test_linalg.TestLinalgCPU testMethod=test_det_cpu_float64>, device = 'cpu', dtype = torch.float64

    @skipCUDAIfNoMagma
    @skipCPUIfNoLapack
    @unittest.skipIf(not TEST_NUMPY, "NumPy not found")
    @dtypes(torch.double)
    def test_det(self, device, dtype):
        tensors = (
            torch.randn((2, 2), device=device, dtype=dtype),
            torch.randn((129, 129), device=device, dtype=dtype),
            torch.randn((3, 52, 52), device=device, dtype=dtype),
            torch.randn((4, 2, 26, 26), device=device, dtype=dtype))
    
    
        ops = (torch.det, torch.Tensor.det,
               torch.linalg.det)
        for t in tensors:
            expected = np.linalg.det(t.cpu().numpy())
            for op in ops:
                actual = op(t)
                self.assertEqual(actual, expected)
    
        # NOTE: det requires a 2D+ tensor
        t = torch.randn(1, device=device, dtype=dtype)
        with self.assertRaises(IndexError):
>           op(t)
E           RuntimeError: Tensor of matrices must have at least 2 dimensions.

test/test_linalg.py:60: RuntimeError
===== short test summary info =====

cc @mruberry @VitalyFedyunin

Metadata

Metadata

Assignees

Labels

module: testsIssues related to tests (not the torch.testing module)triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions