Bug Fix for offload_states API #7050
Merged
+16
−9
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
bugfix for offload_states Signed-off-by: Wei Wu <[email protected]>
U-rara
force-pushed
the
bugfix_reload_states
branch
from
00fa6a4 to
836b55b
Compare
tjruwase
reviewed
Feb 20, 2025
tjruwase
reviewed
Feb 20, 2025
Signed-off-by: Wei Wu <[email protected]>
|
@tjruwase Good suggestion! My previous fix seemed a bit hasty, so I optimized it according to your advice. |
tjruwase
approved these changes
Feb 20, 2025
github-merge-queue
bot
removed this pull request from the merge queue due to no response for status checks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
@fukun07 and I discovered a bug when using the
offload_statesandreload_statesAPIs of the Zero3 optimizer. When using grouped parameters (for example, in weight decay or grouped lr scenarios), the order of the parameters mapping inreload_states(here) does not correspond with the initialization ofself.lp_param_buffer(here), which leads to misaligned parameter loading. This issue was overlooked by the corresponding unit tests (here), so we fixed the bug in our PR and added the corresponding unit tests.