Split parameter offload from z3 #2009
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
tjruwase
requested review from
ShadenSmith,
conglongli,
awan-10,
cli99,
eltonzheng,
minjiaz and
RezaYazdaniAminabadi
as code owners
|
I can confirm that with this PR, I'm able to run CPU-offload on a 6B parameters models with limited RAM (32GB). Usage peaks at 24GB before going down to a constant 20GB. The same test script crashes due to OOM on the last version of DeepSpeed. |
|
Thank you for validating the fix, @sgugger! Tunji, is it good to merge? as quite a few users are impacted by this memory issue and your PR would help to many. Thank you! |
…pSpeed into olruwase/parameter_offload
tjruwase
removed request for
ShadenSmith,
conglongli,
awan-10,
siddharth9820,
cli99,
eltonzheng and
minjiaz
tjruwase
requested review from
duli2012,
yaozhewei,
arashb and
xiaoxiawu-microsoft
as code owners
tjruwase
removed request for
duli2012,
yaozhewei,
arashb and
xiaoxiawu-microsoft
mrwyattii
approved these changes
Jun 15, 2022
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Outdated suggestions cannot be applied.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.
Parameter offloading can live separately from stage 3 optimizer, right?