Skip to content

Move RNN implementations to C++#10481

Closed
apaszke wants to merge 10 commits intopytorch:masterfrom
apaszke:rnn_move
Closed

Move RNN implementations to C++#10481
apaszke wants to merge 10 commits intopytorch:masterfrom
apaszke:rnn_move

Conversation

@apaszke
Copy link
Contributor

@apaszke apaszke commented Aug 13, 2018

This is the first of two changes that are supposed to improve how we handle RNNs in the JIT. They still get traced as PythonOps, but now it will be much easier to actually expose them to the JIT as e.g. aten::lstm, and ignore the Python interpreter entirely. This needs some symbolic adjustments that will be part of a second PR.

Even when we fix symbolics, there will still be a bit of a problem with statefulness of the cuDNN API (we need a mutable cache for the dropout state, but our IR has no way of representing that).

@zdevito @ezyang

ezyang
ezyang reviewed Aug 13, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

@ezyang
Copy link
Contributor

ezyang commented Aug 14, 2018

Haven't finished yet. If you want to merge this AOT and fix up residual comments later I'm OK with that.

@apaszke
Copy link
Contributor Author

apaszke commented Aug 14, 2018

Thanks for a quick review, but I'll wait. I can find unrelated tasks to work on in the meantime.

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang reviewed Aug 14, 2018
View reviewed changes
aten/src/ATen/native/RNN.cpp Outdated

This comment was marked as off-topic.

Sign in to view

This comment was marked as off-topic.

Sign in to view

ezyang
ezyang approved these changes Aug 14, 2018
View reviewed changes
Copy link
Contributor

@ezyang ezyang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cool stuff.

facebook-github-bot
facebook-github-bot reviewed Aug 15, 2018
View reviewed changes
Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

apaszke has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

facebook-github-bot
facebook-github-bot reviewed Aug 15, 2018
View reviewed changes
Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

apaszke has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

zdevito pushed a commit to zdevito/ATen that referenced this pull request Aug 15, 2018
@apaszke @facebook-github-bot
Move RNN implementations to C++ (#10481)
eef4f84 
Summary:
This is the first of two changes that are supposed to improve how we handle RNNs in the JIT. They still get traced as `PythonOp`s, but now it will be much easier to actually expose them to the JIT as e.g. `aten::lstm`, and ignore the Python interpreter entirely. This needs some symbolic adjustments that will be part of a second PR.

Even when we fix symbolics, there will still be a bit of a problem with statefulness of the cuDNN API (we need a mutable cache for the dropout state, but our IR has no way of representing that).

zdevito ezyang
Pull Request resolved: pytorch/pytorch#10481

Reviewed By: ezyang

Differential Revision: D9341113

Pulled By: apaszke

fbshipit-source-id: 0ae30ead72a1b12044b7c12369d11e5ca8ec30b5
@apaszke apaszke deleted the rnn_move branch August 17, 2018 15:54
@goldsborough goldsborough mentioned this pull request Aug 21, 2018
Closed
facebook-github-bot pushed a commit that referenced this pull request Aug 22, 2018
@goldsborough @facebook-github-bot
Create at::linear (#10755)
1068ba6 
Summary:
The optimized code for `linear()` which uses `addmm` when a bias is given was duplicated three times in the ATen and the C++ API. Let's just have `at::linear` and use that everywhere.

apaszke ezyang (who mentioned this in #10481)
Pull Request resolved: #10755

Differential Revision: D9443881

Pulled By: goldsborough

fbshipit-source-id: a64862d1649b5961043d58401625ec267d97d9f3
@mcarilli mcarilli mentioned this pull request Aug 22, 2018
Closed
PenghuiCheng pushed a commit to PenghuiCheng/pytorch that referenced this pull request Sep 11, 2018
@goldsborough @PenghuiCheng
Create at::linear (pytorch#10755)
2cc90e4 
Summary:
The optimized code for `linear()` which uses `addmm` when a bias is given was duplicated three times in the ATen and the C++ API. Let's just have `at::linear` and use that everywhere.

apaszke ezyang (who mentioned this in pytorch#10481)
Pull Request resolved: pytorch#10755

Differential Revision: D9443881

Pulled By: goldsborough

fbshipit-source-id: a64862d1649b5961043d58401625ec267d97d9f3
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@ezyang ezyang ezyang approved these changes

@facebook-github-bot facebook-github-bot facebook-github-bot left review comments

@colesbury colesbury Awaiting requested review from colesbury

@gchanan gchanan Awaiting requested review from gchanan

@soumith soumith Awaiting requested review from soumith

@zdevito zdevito Awaiting requested review from zdevito

Assignees

No one assigned

Labels

open source

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@apaszke @ezyang @facebook-github-bot