Skip to content

Enable CUDA usage of vesin in MetatomicCalculator #571

Enable CUDA usage of vesin in MetatomicCalculator

Enable CUDA usage of vesin in MetatomicCalculator #571

Triggered via pull request February 13, 2026 16:49
Status Cancelled
Total duration 2m 15s
Artifacts

torch-tests.yml

on: pull_request
Matrix: tests
Fit to window
Zoom out
Zoom in

Annotations

11 errors and 5 notices
ubuntu-24.04 / Python 3.10 / Torch 2.1
Canceling since a higher priority waiting request for torch-tests-refs/pull/157/merge exists
ubuntu-24.04 / Python 3.10 / Torch 2.1
The operation was canceled.
windows-2022 / Python 3.13 / Torch 2.10
Canceling since a higher priority waiting request for torch-tests-refs/pull/157/merge exists
windows-2022 / Python 3.13 / Torch 2.10
The operation was canceled.
ubuntu-24.04 / Python 3.10 / Torch 2.10
Canceling since a higher priority waiting request for torch-tests-refs/pull/157/merge exists
ubuntu-24.04 / Python 3.10 / Torch 2.10
The operation was canceled.
ubuntu-24.04 / Python 3.13 / Torch 2.10
Canceling since a higher priority waiting request for torch-tests-refs/pull/157/merge exists
ubuntu-24.04 / Python 3.13 / Torch 2.10
The operation was canceled.
macos-15 / Python 3.13 / Torch 2.10
Canceling since a higher priority waiting request for torch-tests-refs/pull/157/merge exists
macos-15 / Python 3.13 / Torch 2.10
The operation was canceled.
TorchScript tests
Canceling since a higher priority waiting request for torch-tests-refs/pull/157/merge exists
sccache stats
100% - 6 hits, 0 misses, 0 errors
sccache stats
0% - 0 hits, 0 misses, 0 errors
sccache stats
93% - 13 hits, 1 misses, 0 errors
sccache stats
100% - 12 hits, 0 misses, 0 errors
sccache stats
0% - 0 hits, 0 misses, 0 errors