Enable test_linalg_solve_triangular_large (pytorch#96182)

PR to see if test fails after removing skip line

Fixes pytorch#70111
Pull Request resolved: pytorch#96182
Approved by: https://github.com/lezcano

Author: iremyux

test/test_linalg.py

@@ -4046,7 +4046,6 @@ def test_linalg_solve_triangular(self, device, dtype):
     @onlyCUDA
     @skipCUDAIfNoMagma  # Magma needed for the PLU decomposition
     @skipCUDAIfRocm  # There is a memory access bug in rocBLAS in the (non-batched) solve_triangular
-    @skipCUDAVersionIn([(11, 3), (11, 6), (11, 7)])  # Tracked in https://github.com/pytorch/pytorch/issues/70111
     @dtypes(*floating_and_complex_types())
     @precisionOverride({torch.float32: 1e-2, torch.complex64: 1e-2,
                         torch.float64: 1e-8, torch.complex128: 1e-8})