Skip to content

Conversation

albertvillanova
Copy link
Member

Implement specific cache directories (datasets, metrics and modules) per test function call.

Currently, the cache directories are set within the temporary test directory, but they are shared across all test function calls.

This PR implements specific cache directories for each test function call, so that tests are atomic and there are no side effects.

@albertvillanova
Copy link
Member Author

@lhoestq, I think this reaches some memory limit on Linux instances... (?)

@lhoestq
Copy link
Member

lhoestq commented Apr 21, 2021

It looks like the comet metric test fails because it tries to load a model in memory.
In the tests I think we have patch_comet that mocks the model download + inference. Not sure why it didn't work though.
I can take a look tomorrow (this afternoon is the pytorch ecosystem day)

@albertvillanova
Copy link
Member Author

albertvillanova commented Apr 21, 2021

@lhoestq thanks for the hint: I'm going to have a look at that mock... ;)

@albertvillanova
Copy link
Member Author

@lhoestq finally I did not find out why the mock is not used... If you can give me some other hint tomorrow...

@albertvillanova albertvillanova added this to the 1.7 milestone Apr 22, 2021
@albertvillanova albertvillanova modified the milestones: 1.7, 1.8 May 31, 2021
@albertvillanova albertvillanova modified the milestones: 1.8, 1.9 Jun 8, 2021
@albertvillanova albertvillanova modified the milestones: 1.9, 1.10 Jul 9, 2021
@albertvillanova albertvillanova modified the milestones: 1.10, 1.11 Jul 21, 2021
@albertvillanova albertvillanova modified the milestones: 1.11, 1.12 Sep 2, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants