[FSDP2] Added test to show rank 0 CPU full state dict flow #124741
Closed
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
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/124741
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (1 Unrelated Failure)As of commit 6726ed3 with merge base c82fcb7 ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
cc mrshenli pritamdamania87 zhaojuanmao satgera rohan-varma gqchen aazzolini osalpekar jiayisuse H-Huang kwen2501 penguinwu fegin XilunWu wanchaol fduwjj wz337 tianyu-l wconstab yf225 chauhang d4l3k [ghstack-poisoned]
This PR adds a unit test to show how we can convert FSDP2 GPU sharded state dicts to a CPU full state dict on rank 0. cc mrshenli pritamdamania87 zhaojuanmao satgera rohan-varma gqchen aazzolini osalpekar jiayisuse H-Huang kwen2501 penguinwu fegin XilunWu wanchaol fduwjj wz337 tianyu-l wconstab yf225 chauhang d4l3k [ghstack-poisoned]
This was referenced Apr 23, 2024
wanchaol
approved these changes
Apr 23, 2024
| torch.manual_seed(42) | ||
| ref_model = Transformer(model_args).cuda() | ||
| for param in ref_model.parameters(): | ||
| torch.distributed.broadcast(param.detach(), src=0) |
There was a problem hiding this comment.
do we have some logic in fully_shard to broadcast the initial tensors?
There was a problem hiding this comment.
We do not right now. This broadcast here is actually because we are using MTPG, so I think that there is no way to ensure the same starting reference model on all ranks (since seed is per process, not per thread).
wz337
approved these changes
Apr 23, 2024
|
@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 |
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 24, 2024
This PR makes sure to construct the `DeviceMesh`'s `mesh` tensor on CPU device in `init_device_mesh()`. This means that we can call `init_device_mesh()` under meta-device context and still construct the correct `mesh` tensor. Pull Request resolved: #124767 Approved by: https://github.com/wz337 ghstack dependencies: #124651, #124741
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 24, 2024
awgu
pushed a commit
to awgu/pytorch
that referenced
this pull request
Apr 24, 2024
ghstack-source-id: 1010a27 Pull Request resolved: pytorch#124741
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 24, 2024
This PR adds a `DeviceMesh.from_group()` static method to convert an existing process group to a device mesh. Motivation: We need `DeviceMesh.from_group()` to allow FSDP2 to interoperate with distributed libraries that do not use `DeviceMesh` for all parallelisms. Pull Request resolved: #124787 Approved by: https://github.com/wanchaol ghstack dependencies: #124651, #124741, #124767, #124768, #124780
This was referenced Apr 25, 2024
alat-rights
pushed a commit
to alat-rights/pytorch
that referenced
this pull request
Apr 26, 2024
This PR adds a `DeviceMesh.from_group()` static method to convert an existing process group to a device mesh. Motivation: We need `DeviceMesh.from_group()` to allow FSDP2 to interoperate with distributed libraries that do not use `DeviceMesh` for all parallelisms. Pull Request resolved: pytorch#124787 Approved by: https://github.com/wanchaol ghstack dependencies: pytorch#124651, pytorch#124741, pytorch#124767, pytorch#124768, pytorch#124780
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 29, 2024
This PR renames the `FSDP` class to `FSDPModule`. This is a BC breaking change. The rationale is that `FSDPModule` is more descriptive since `fully_shard` is a module-level API (applied to a `module` arg), so the `FSDP` class will always correspond to a module. Also, users commonly import `FullyShardedDataParallel` as `FSDP`, so this can help avoid some name conflict in some cases. Pull Request resolved: #124955 Approved by: https://github.com/wanchaol, https://github.com/wconstab ghstack dependencies: #124651, #124741, #124767, #124768, #124780, #124787
This was referenced Apr 29, 2024
pytorch-bot bot
pushed a commit
that referenced
this pull request
May 3, 2024
This PR adds a unit test to show how we can convert FSDP2 GPU sharded state dicts to a CPU full state dict on rank 0. Pull Request resolved: #124741 Approved by: https://github.com/wanchaol, https://github.com/wz337 ghstack dependencies: #124651
petrex
pushed a commit
to petrex/pytorch
that referenced
this pull request
May 3, 2024
This PR makes sure to construct the `DeviceMesh`'s `mesh` tensor on CPU device in `init_device_mesh()`. This means that we can call `init_device_mesh()` under meta-device context and still construct the correct `mesh` tensor. Pull Request resolved: pytorch#124767 Approved by: https://github.com/wz337 ghstack dependencies: pytorch#124651, pytorch#124741
petrex
pushed a commit
to petrex/pytorch
that referenced
this pull request
May 3, 2024
Pull Request resolved: pytorch#124768 Approved by: https://github.com/wz337 ghstack dependencies: pytorch#124651, pytorch#124741, pytorch#124767
petrex
pushed a commit
to petrex/pytorch
that referenced
this pull request
May 3, 2024
This PR adds a `DeviceMesh.from_group()` static method to convert an existing process group to a device mesh. Motivation: We need `DeviceMesh.from_group()` to allow FSDP2 to interoperate with distributed libraries that do not use `DeviceMesh` for all parallelisms. Pull Request resolved: pytorch#124787 Approved by: https://github.com/wanchaol ghstack dependencies: pytorch#124651, pytorch#124741, pytorch#124767, pytorch#124768, pytorch#124780
pytorch-bot bot
pushed a commit
that referenced
this pull request
May 3, 2024
This PR renames the `FSDP` class to `FSDPModule`. This is a BC breaking change. The rationale is that `FSDPModule` is more descriptive since `fully_shard` is a module-level API (applied to a `module` arg), so the `FSDP` class will always correspond to a module. Also, users commonly import `FullyShardedDataParallel` as `FSDP`, so this can help avoid some name conflict in some cases. Pull Request resolved: #124955 Approved by: https://github.com/wanchaol, https://github.com/wconstab ghstack dependencies: #124651, #124741, #124767, #124768, #124780, #124787
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.
Stack from ghstack (oldest at bottom):
DeviceMesh.from_group()#124787.to(cpu)#124768This PR adds a unit test to show how we can convert FSDP2 GPU sharded state dicts to a CPU full state dict on rank 0.
cc @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @osalpekar @jiayisuse @H-Huang @kwen2501 @penguinwu @fegin @XilunWu @wanchaol @fduwjj @wz337 @tianyu-l @wconstab @yf225 @chauhang @d4l3k