Support submodule wrapping for FSDP2 according to model definition (with _fsdp_wrap and fsdp_wrap_fn)
#3826
Conversation
This file contains hidden or 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
dakinggg
reviewed
Apr 18, 2025
rithwik-db
force-pushed
the
add_fsdp_wrapping
branch
from
5c7e431 to
9c2c52d
Compare
rithwik-db
changed the title
_fsdp_wrap for FSDP2_fsdp_wrap for FSDP2
dakinggg
reviewed
Apr 23, 2025
rithwik-db
force-pushed
the
add_fsdp_wrapping
branch
from
c5fccef to
d4f5ba5
Compare
dakinggg
reviewed
Apr 23, 2025
rithwik-db
force-pushed
the
add_fsdp_wrapping
branch
from
0e4564e to
a79f458
Compare
dakinggg
reviewed
Apr 24, 2025
dakinggg
reviewed
Apr 24, 2025
rithwik-db
force-pushed
the
add_fsdp_wrapping
branch
from
5fd764b to
345392d
Compare
bowenyang008
approved these changes
Apr 28, 2025
dakinggg
approved these changes
Apr 28, 2025
rithwik-db
force-pushed
the
add_fsdp_wrapping
branch
from
8ea27f1 to
02e818e
Compare
rithwik-db
changed the title
_fsdp_wrap for FSDP2_fsdp_wrap and fsdp_wrap_fn)
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Outdated suggestions cannot be applied.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.
What does this PR do?
Modified existing code to support
_fsdp_wrapand_fsdp_wrap_fnsimilar to how FSDP1 handles it (with the slight caveat that we allowNoneto be a valid input for skipping the current module and checking descendants).Reworked module wrapping so that we do a recursive check where we pre-order check the legalization and validation of params to make sure nothing that shouldn't be weight tied is actually tied and then we do a post-order fully_shard on valid parameters.