[Gradient Compression] Surface C++ comm hooks to Python API as built-in comm hooks #47270
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
…in comm hooks This is almost same as #46959, except that in caffe2/torch/nn/parallel/distributed.py, BuiltinCommHookType is imported conditionally, only when dist.is_available(). Otherwise, this Python enum type defined in caffe2/torch/scrc/distributed/c10d/init.cpp cannot be imported, which is similar to another enum type ReduceOp defined in the same file. See #47153 To review the diff on top of #46959, compare V1 vs Latest. Main Changes in V1 (#46959): 1. Implemented the Pybind part. 2. In the reducer, once the builtin_comm_hook_type is set, a c++ comm hook instance will be created in Reducer::autograd_hook. 3. Added unit tests for the builit-in comm hooks. Original PR issue: C++ DDP Communication Hook #46348 Differential Revision: [D24700959](https://our.internmc.facebook.com/intern/diff/D24700959/) [ghstack-poisoned]
wayi1
requested review from
apaszke,
mingzhe09088,
mrshenli,
pietern,
pritamdamania87,
rohan-varma and
zhaojuanmao
as code owners
facebook-github-bot
added
cla signed
oncall: distributed
wayi1
pushed a commit
that referenced
this pull request
Nov 3, 2020
…in comm hooks This is almost same as #46959, except that in caffe2/torch/nn/parallel/distributed.py, BuiltinCommHookType is imported conditionally, only when dist.is_available(). Otherwise, this Python enum type defined in caffe2/torch/scrc/distributed/c10d/init.cpp cannot be imported, which is similar to another enum type ReduceOp defined in the same file. See #47153 To review the diff on top of #46959, compare V1 vs Latest. Main Changes in V1 (#46959): 1. Implemented the Pybind part. 2. In the reducer, once the builtin_comm_hook_type is set, a c++ comm hook instance will be created in Reducer::autograd_hook. 3. Added unit tests for the builit-in comm hooks. Original PR issue: C++ DDP Communication Hook #46348 Differential Revision: [D24700959](https://our.internmc.facebook.com/intern/diff/D24700959/) ghstack-source-id: 115753518 Pull Request resolved: #47270
Codecov Report
@@ Coverage Diff @@
## gh/SciPioneer/24/base #47270 +/- ##
=========================================================
- Coverage 60.84% 60.83% -0.01%
=========================================================
Files 2749 2749
Lines 254114 254144 +30
=========================================================
+ Hits 154605 154618 +13
- Misses 99509 99526 +17 |
…I as built-in comm hooks" This is almost same as #46959, except that in caffe2/torch/nn/parallel/distributed.py, BuiltinCommHookType is imported conditionally, only when dist.is_available(). Otherwise, this Python enum type defined in caffe2/torch/scrc/distributed/c10d/init.cpp cannot be imported, which is similar to another enum type ReduceOp defined in the same file. See #47153 Main Changes in #46959: 1. Implemented the Pybind part. 2. In the reducer, once the builtin_comm_hook_type is set, a c++ comm hook instance will be created in Reducer::autograd_hook. 3. Added unit tests for the builit-in comm hooks. Original PR issue: C++ DDP Communication Hook #46348 Differential Revision: [D24700959](https://our.internmc.facebook.com/intern/diff/D24700959/) [ghstack-poisoned]
wayi1
pushed a commit
that referenced
this pull request
Nov 3, 2020
…in comm hooks Pull Request resolved: #47270 This is almost same as #46959, except that in caffe2/torch/nn/parallel/distributed.py, BuiltinCommHookType should be imported conditionally, only when dist.is_available(). Otherwise, this Python enum type defined in caffe2/torch/scrc/distributed/c10d/init.cpp cannot be imported. See #47153 I tried to follow another enum type enum type ReduceOp defined in the same file, but did not work, because the C++ enum class is defined torch/lib/c10d library, but BuiltinCommHookType is defined in torch/csrc/distributed library. These two libraries are compiled in two different ways. To avoid adding typing to distributed package, which can be a new project, I simply removed the arg type of BuiltinCommHookType in this file. To review the diff on top of #46959, compare V1 vs Latest: https://www.internalfb.com/diff/D24700959?src_version_fbid=270445741055617 Main Changes in V1 (#46959): 1. Implemented the Pybind part. 2. In the reducer, once the builtin_comm_hook_type is set, a c++ comm hook instance will be created in Reducer::autograd_hook. 3. Added unit tests for the builit-in comm hooks. Original PR issue: C++ DDP Communication Hook #46348 ghstack-source-id: 115783237 Differential Revision: [D24700959](https://our.internmc.facebook.com/intern/diff/D24700959/)
|
Just a heads-up, there might be conflict with #47309. If so, please coordinate landing with @gmagogsfm |
mrshenli
approved these changes
Nov 3, 2020
|
This pull request has been merged in f91fcef. |
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:
This is almost same as #46959, except that in caffe2/torch/nn/parallel/distributed.py, BuiltinCommHookType should be imported conditionally, only when dist.is_available(). Otherwise, this Python enum type defined in caffe2/torch/scrc/distributed/c10d/init.cpp cannot be imported. See #47153
I tried to follow another enum type enum type ReduceOp defined in the same file, but did not work, because the C++ enum class is defined torch/lib/c10d library, but BuiltinCommHookType is defined in torch/csrc/distributed library. These two libraries are compiled in two different ways.
To avoid adding typing to distributed package, which can be a new project, I simply removed the arg type of BuiltinCommHookType in this file.
Main Changes in #46959:
Original PR issue: C++ DDP Communication Hook #46348
Differential Revision: D24700959