add runtime check for PyTorch 2.9.1 + CuDNN < 9.15 to prevent Conv3d performance issues (#14119)
This commit is contained in:
2
.github/workflows/pr-test.yml
vendored
2
.github/workflows/pr-test.yml
vendored
@@ -453,8 +453,8 @@ jobs:
|
||||
|
||||
- name: Install dependencies
|
||||
run: |
|
||||
CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
|
||||
pip install "bitsandbytes>=0.44.0"
|
||||
CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
|
||||
- name: Run test
|
||||
timeout-minutes: 30
|
||||
run: |
|
||||
|
||||
Reference in New Issue
Block a user