OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-07 12:21:27 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity

[CUDA][cuBLAS] Bump test_cublas_baddbmm_large_input tolerances (#117889)

Browse Source 
Unfortunate that the current `rtol=1e-5` hits a literal 1 / 1000000 mismatch (`rtol=1.04e-5`) on L40.

CC @ptrblck

Pull Request resolved: https://github.com/pytorch/pytorch/pull/117889
Approved by: https://github.com/atalman
This commit is contained in:
eqy 2024-02-27 19:05:20 +00:00 committed by PyTorch MergeBot
parent 5b5c167adc
commit 65efece3a4
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
test/test_matmul_cuda.py
View File

@ -150,7 +150,7 @@ class TestMatmulCuda(TestCase):
@onlyCUDA
@unittest.skipIf(IS_JETSON, "Too large for Jetson")
@toleranceOverride({torch.float32: xtol(atol=1e-5, rtol=1e-5)})
@toleranceOverride({torch.float32: xtol(atol=1e-5, rtol=1.1e-5)})
@dtypes(*([torch.float32, torch.float16] +
[torch.bfloat16] if TEST_WITH_ROCM or SM53OrLater else []))
@parametrize(