This repository was archived by the owner on Nov 3, 2023. It is now read-only.
Bug fix: Multi-task with datatype train:stream #2788
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
added 3 commits
June 23, 2020 15:24
|
Your PR contains a change to a task. Please paste the results of the following command into a comment: python tests/datatests/test_new_tasks.py |
stephenroller
approved these changes
Jun 24, 2020
|
Why aren't tests running tho |
@klshuster 's idea |
klshuster
approved these changes
Jun 24, 2020
parlai/utils/data.py
Outdated
| """ | ||
|
|
||
|
|
||
| def is_training(datatype: str) -> bool: |
|
tbh I'm surprised that chunky teacher doesn't hang. |
|
also since tests aren't running here, maybe you can run tests locally? just to check |
|
Filed a bug with CircleCI. None of Emily's PRs will run tests until they fix something. |
klshuster
approved these changes
Jun 24, 2020
parlai/utils/data.py
Outdated
| parlai datatype | ||
|
|
||
| :return should_shuffle: | ||
| given datatype, return whether we should cycle |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Patch description
@klshuster found a bug where if we multi-task with
-dt train:stream, the shorter teacher will only complete 1 epoch. This is fixed by: (1) changing the condition we check for FixedDialogTeacher settingepoch_doneand (2) changingself.randomfor multiworld. In the multiworld setting, we want to randomly choose between subworlds when training, even if the examples in the subworld may be ordered (as intrain:stream).I wrote a test that fails before these changes and now passes.