-
Notifications
You must be signed in to change notification settings - Fork 24.8k
Open
Labels
high prioritymodule: ciRelated to continuous integrationRelated to continuous integrationmodule: cudaRelated to torch.cuda, and CUDA support in generalRelated to torch.cuda, and CUDA support in generaltriagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module
Description
We are working on PR to move our Inductor CUDA 12.4 -> 12.6
This is pull request:
#148612
We do see 2 test failures and 1 pass
fail_accuracy:
timm_efficientnet,fail_accuracy,7
crossvit_9_240,fail_accuracy,7
pass:
tinynet_a,pass,6
Please note: this looks like an improvement, it was turned off with this PR: #123475
Workflows:
29ec07b#diff-edec9c9b528f9ee5f7113d88ce1e776897a63c5ac0260c55926a570a62a0a36b
and
https://github.com/pytorch/pytorch/actions/runs/13693492178/job/38295168222
Versions
2.7.0 nightly
cc @ptrblck @msaroufim @eqy @seemethere @malfet @pytorch/pytorch-dev-infra @ezyang @gchanan @zou3519 @kadeng @chauhang @penguinwu
tinglvv, eqy and FireballDWF
Metadata
Metadata
Assignees
Labels
high prioritymodule: ciRelated to continuous integrationRelated to continuous integrationmodule: cudaRelated to torch.cuda, and CUDA support in generalRelated to torch.cuda, and CUDA support in generaltriagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module