search

aimhubio / aim

Aim 💫 — An easy-to-use & supercharged open-source experiment tracker.
https://aimstack.io
Apache License 2.0
4.93k stars 298 forks source link

[fix] Force reacquire lock in torch lightning #3090

Closed inc0 closed 5 months ago

inc0 commented 5 months ago

When running testing on a Run, this line failed with error similar to #2999 . It was quite hard to debug, but it seems this is the culprit. It appears that Run was already locked and it failed with traceback when trying to run pytorch lightning training+testing run

inc0 commented 5 months ago

Thanks @alberttorosyan for review! Can you take a look at the CI failure please? It doesn't seem to be related to the PR.

alberttorosyan commented 5 months ago

Thanks @alberttorosyan for review! Can you take a look at the CI failure please? It doesn't seem to be related to the PR.

Yes, the unit-tests passed after re-run. Thank you for your contribution 🙌