You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the CI runs once a week. With the upcoming DL_edition update the runs become much heavier. So until we have a more fine-grained checking mechanism, we can probably stretch that to once a month.
The text was updated successfully, but these errors were encountered:
Once a month seems a bit too long for this repo that is being used quite a lot for teaching, in my opinion. Maybe every two weeks? But up to you @AndreaVolkamer.
If the DL talktorials take too long in the CI, can you think of ways to make this work more elegantly during testing (we do have quite a few notebooks where we adapted the code in a way that it will be less heavy during CI runs --- if you point me to the code in question, I can think about it; nothing that needs to be done for the DL release).
Typically, some of the 4 jobs fail, due to online resources not being reachable for whatever reason. In that case you have to re-start the jobs a couple of times. So each time there is a pretty high probability of manual intervention being necessary.
Maybe it is also possible to add a retrial scheme for notebooks that use such resources (T001, T018...)?
Currently, the CI runs once a week. With the upcoming DL_edition update the runs become much heavier. So until we have a more fine-grained checking mechanism, we can probably stretch that to once a month.
The text was updated successfully, but these errors were encountered: