Pinned torch==2.2.2
breaks CI due to numpy>=2.0
#812
Labels
dependencies 🔗
Involves external dependency libraries
type: ci 🔧
Related to Continuous Integration workflows
Milestone
What should we add?
The CI tests involving
TorchConnector
fail withhttps://github.com/qiskit-community/qiskit-machine-learning/actions/runs/9753721897/job/26919448830#step:8:2473
because version 2.2.2 of PyTorch, pinned in #796, is no longer compatible with Numpy following the v. 2.0 release.
Proposed solution
Release the
torch==2.2.2
constraint to install the latest Numpy 2.0-compatible version (>=2.3.1), keeping in mind the UTF-8 issue in PyTorch 2.3.0, #797 and pytorch/pytorch#124897, that was affecting the CI tests on Windows platform.The text was updated successfully, but these errors were encountered: