This repository was archived by the owner on Nov 22, 2022. It is now read-only.
auto resume from latest checkpoint for FBLearner retry job #1046
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
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D17874151 |
hudeven
added a commit
to hudeven/pytext
that referenced
this pull request
Oct 16, 2019
…esearch#1046) Summary: Pull Request resolved: facebookresearch#1046 ## Problem: When a job fail at epoch X, FBLearner will retry it 3 times, from scratch(epoch1) It wastes time and resources ## Solution: 1. The "retry" shares the same FBLearner job_id 2. we save all checkpoints at "manifold://pytext_training/tree/jobs/<JOB_ID>/checkpoint-<EPOCH>" 3. so there is a mapping (job_id --> latest checkpoint) 4. before training start, we check if there are already checkpoints saved for the current job_id. Yes: load the latest checkpoint and resume training No: start training from scratch 5. add config "auto_resume_from_snapshot" to turn this feature on/off. Reviewed By: mwu1993 Differential Revision: D17874151 fbshipit-source-id: 6778e68e90523e982c141b76e916cc7d3038de8f
hudeven
force-pushed
the
export-D17874151
branch
from
ae495fe to
904c3f7
Compare
|
This pull request was exported from Phabricator. Differential Revision: D17874151 |
…esearch#1046) Summary: Pull Request resolved: facebookresearch#1046 ## Problem: When a job fail at epoch X, FBLearner will retry it 3 times, from scratch(epoch1) It wastes time and resources ## Solution: 1. The "retry" shares the same FBLearner job_id 2. we save all checkpoints at "manifold://pytext_training/tree/jobs/<JOB_ID>/checkpoint-<EPOCH>" 3. so there is a mapping (job_id --> latest checkpoint) 4. before training start, we check if there are already checkpoints saved for the current job_id. Yes: load the latest checkpoint and resume training No: start training from scratch 5. add config "auto_resume_from_snapshot" to turn this feature on/off. Reviewed By: mwu1993 Differential Revision: D17874151 fbshipit-source-id: ec17cbd1a5f5022b47e2371e9063c738b67d9018
hudeven
force-pushed
the
export-D17874151
branch
from
904c3f7 to
3490927
Compare
|
This pull request was exported from Phabricator. Differential Revision: D17874151 |
|
This pull request has been merged in 8f93ce1. |
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.
Summary:
Problem:
When a job fail at epoch X, FBLearner will retry it 3 times, from scratch(epoch1)
It wastes time and resources
Solution:
Yes: load the latest checkpoint and resume training
No: start training from scratch
Differential Revision: D17874151