[FSDP] Propagate requires_grad attribute to unsharded params #109892
Closed
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
edpizzi
requested review from
mrshenli,
zhaojuanmao,
rohan-varma,
H-Huang,
awgu,
kwen2501,
wanchaol,
fegin,
fduwjj,
wz337,
kiukchung and
d4l3k
as code owners
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/109892
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 168c5d2 with merge base 92de1d3 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
|
pytorch-bot
bot
added
the
release notes: distributed (fsdp)
|
This pull request was exported from Phabricator. Differential Revision: D49517155 |
edpizzi
added a commit
to edpizzi/pytorch
that referenced
this pull request
Sep 22, 2023
…#109892) Summary: This preserves `requires_grad` in the case where all parameters within a `FlatParameter` have the same `requires_grad` value. Currently, unsharded parameters have `requires_grad=True` in some cases where the `FlatParameter` and all original parameters have `requires_grad=False`. This could be extended to support `FlatParameters` with a mix of `requires_grad` states by extending `ParamInfo` to capture `requires_grad` for each parameter. Test Plan: test added Reviewed By: awgu Differential Revision: D49517155
edpizzi
force-pushed
the
export-D49517155
branch
from
2a94937
to
9bc05e0
Compare
…#109892) Summary: This preserves `requires_grad` in the case where all parameters within a `FlatParameter` have the same `requires_grad` value. Currently, unsharded parameters have `requires_grad=True` in some cases where the `FlatParameter` and all original parameters have `requires_grad=False`. This could be extended to support `FlatParameters` with a mix of `requires_grad` states by extending `ParamInfo` to capture `requires_grad` for each parameter. Test Plan: test added Reviewed By: awgu Differential Revision: D49517155
|
This pull request was exported from Phabricator. Differential Revision: D49517155 |
edpizzi
force-pushed
the
export-D49517155
branch
from
9bc05e0
to
168c5d2
Compare
|
This pull request was exported from Phabricator. Differential Revision: D49517155 |
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: 1 jobs have failed, first few of them are: trunk / linux-focal-rocm5.6-py3.8 / test (default, 2, 3, linux.rocm.gpu) Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
edpizzi
added a commit
to edpizzi/d2go
that referenced
this pull request
Sep 25, 2023
…checkpoints Summary: EMA can be configured to exclude frozen (`requires_grad=False`) parameters and buffers, reducing memory use and checkpoint size. However `FULL_STATE_DICT` FSDP + EMA checkpoints construct an inner `EMAState` after unsharding FSDP parameters. This inner `EMAState` uses default `include_frozen` and `include_buffers` settings, resulting in checkpoints containing frozen parameters and buffers regardless of settings. Propagate `include_frozen` and `include_buffers` settings to the inner `EMAState` when gathering `FULL_STATE_DICT` FSDP EMA state. This change only affects frozen parameters with a parallel fix to PyTorch FSDP to propagate `requires_grad` across parameter sharding/unsharding: pytorch/pytorch#109892. Differential Revision: D49517178 fbshipit-source-id: fefc9bb898a93c1746dad02007ca5b3384b0fcd6
facebook-github-bot
pushed a commit
to facebookresearch/d2go
that referenced
this pull request
Sep 25, 2023
…checkpoints Summary: Pull Request resolved: #620 EMA can be configured to exclude frozen (`requires_grad=False`) parameters and buffers, reducing memory use and checkpoint size. However `FULL_STATE_DICT` FSDP + EMA checkpoints construct an inner `EMAState` after unsharding FSDP parameters. This inner `EMAState` uses default `include_frozen` and `include_buffers` settings, resulting in checkpoints containing frozen parameters and buffers regardless of settings. Propagate `include_frozen` and `include_buffers` settings to the inner `EMAState` when gathering `FULL_STATE_DICT` FSDP EMA state. This change only affects frozen parameters with a parallel fix to PyTorch FSDP to propagate `requires_grad` across parameter sharding/unsharding: pytorch/pytorch#109892. Reviewed By: daveboat Differential Revision: D49517178 fbshipit-source-id: 0fe159dcec9ec1f2c456ae2ee7798681e7536249
ringohoffman
pushed a commit
to ringohoffman/pytorch
that referenced
this pull request
Sep 27, 2023
…#109892) Summary: This preserves `requires_grad` in the case where all parameters within a `FlatParameter` have the same `requires_grad` value. Currently, unsharded parameters have `requires_grad=True` in some cases where the `FlatParameter` and all original parameters have `requires_grad=False`. This could be extended to support `FlatParameters` with a mix of `requires_grad` states by extending `ParamInfo` to capture `requires_grad` for each parameter. Test Plan: test added Differential Revision: D49517155 Pull Request resolved: pytorch#109892 Approved by: https://github.com/awgu
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.
Summary:
This preserves
requires_gradin the case where all parameters within aFlatParameterhave the samerequires_gradvalue.Currently, unsharded parameters have
requires_grad=Truein some cases where theFlatParameterand all original parameters haverequires_grad=False.This could be extended to support
FlatParameterswith a mix ofrequires_gradstates by extendingParamInfoto capturerequires_gradfor each parameter.Test Plan: test added
Differential Revision: D49517155