More double backwards support for pooling, unpooling, padding #2516
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
|
looks like a python2 issue. |
apaszke
reviewed
Aug 24, 2017
torch/nn/_functions/thnn/auto.py
Outdated
| @@ -333,4 +332,6 @@ def default_double_backwards_fn(ctx, *grad_params): | |||
| _all_functions.append(backward_cls) | |||
|
|
|||
|
|
|||
| from .auto_double_backwards import double_backwards_fns | |||
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
| @@ -106,6 +106,27 @@ def logsoftmax_double_backwards(ctx, ggI): | |||
| return gI, ggO, None, None, None, None | |||
|
|
|||
|
|
|||
| def reflectionpad2d_double_backwards(ctx, ggI): | |||
| gI = None | |||
| ggO = torch.nn._functions.thnn.auto.ReflectionPad2d.apply(ggI, *ctx.additional_args) | |||
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
houseroad
added a commit
to houseroad/pytorch
that referenced
this pull request
Jan 14, 2020
…740f8f (pytorch#32125) Summary: Pull Request resolved: pytorch#32125 Previous import was 57ebc587fcf3913b4be93653b0dd58c686447298 Included changes: - **[65020daa](onnx/onnx@65020daa)**: better error message for undefined inputs (pytorch#2540) <Yuxin Wu> - **[8afff0e9](onnx/onnx@8afff0e9)**: bump ORT version (pytorch#2538) <Lu Fang> - **[3d9ca57e](onnx/onnx@3d9ca57e)**: fix name of directory (pytorch#2537) <Prasanth Pulavarthi> - **[df8fa2c9](onnx/onnx@df8fa2c9)**: Repository guidelines (pytorch#2539) <Prasanth Pulavarthi> - **[49cc2f02](onnx/onnx@49cc2f02)**: Update CircleCI job to use Python3.6 (pytorch#2527) <bddppq> - **[25ff79a4](onnx/onnx@25ff79a4)**: Fix wrong model version, it's not 12 (the onnx_opset_version()), not 11 (the opset version of the latest stable), but 10 (pytorch#2478) <daquexian> - **[7cebaed5](onnx/onnx@7cebaed5)**: Fix Windows py3.5 CI (pytorch#2529) <bddppq> - **[eddae00e](onnx/onnx@eddae00e)**: Correct the order of arguments of InferShapes (pytorch#2500) <Shinichiro Hamaji> - **[41b5afe6](onnx/onnx@41b5afe6)**: Include <ostream> in common/status.h (pytorch#2519) <Casey Carter> - **[423f1977](onnx/onnx@423f1977)**: add 8 bit support to maxpool op (pytorch#2510) <Ashwini Khade> - **[78593c2f](onnx/onnx@78593c2f)**: add 8 bit support to reducemin and reducemax ops (pytorch#2516) <Ashwini Khade> Test Plan: cont build Differential Revision: D19380034 fbshipit-source-id: 1ea677eed6779d2b3f8e4683225ba856c68159cd
facebook-github-bot
pushed a commit
that referenced
this pull request
Jan 14, 2020
…740f8f (#32125) Summary: Pull Request resolved: #32125 Previous import was 57ebc587fcf3913b4be93653b0dd58c686447298 Included changes: - **[65020daa](onnx/onnx@65020daa)**: better error message for undefined inputs (#2540) <Yuxin Wu> - **[8afff0e9](onnx/onnx@8afff0e9)**: bump ORT version (#2538) <Lu Fang> - **[3d9ca57e](onnx/onnx@3d9ca57e)**: fix name of directory (#2537) <Prasanth Pulavarthi> - **[df8fa2c9](onnx/onnx@df8fa2c9)**: Repository guidelines (#2539) <Prasanth Pulavarthi> - **[49cc2f02](onnx/onnx@49cc2f02)**: Update CircleCI job to use Python3.6 (#2527) <bddppq> - **[25ff79a4](onnx/onnx@25ff79a4)**: Fix wrong model version, it's not 12 (the onnx_opset_version()), not 11 (the opset version of the latest stable), but 10 (#2478) <daquexian> - **[7cebaed5](onnx/onnx@7cebaed5)**: Fix Windows py3.5 CI (#2529) <bddppq> - **[eddae00e](onnx/onnx@eddae00e)**: Correct the order of arguments of InferShapes (#2500) <Shinichiro Hamaji> - **[41b5afe6](onnx/onnx@41b5afe6)**: Include <ostream> in common/status.h (#2519) <Casey Carter> - **[423f1977](onnx/onnx@423f1977)**: add 8 bit support to maxpool op (#2510) <Ashwini Khade> - **[78593c2f](onnx/onnx@78593c2f)**: add 8 bit support to reducemin and reducemax ops (#2516) <Ashwini Khade> Test Plan: cont build Reviewed By: benoitsteiner Differential Revision: D19380034 fbshipit-source-id: ddce8450864a611773b2a32e2f0254c9bb6b6906
wuhuikx
pushed a commit
to wuhuikx/pytorch
that referenced
this pull request
Jan 30, 2020
…740f8f (pytorch#32125) Summary: Pull Request resolved: pytorch#32125 Previous import was 57ebc587fcf3913b4be93653b0dd58c686447298 Included changes: - **[65020daa](onnx/onnx@65020daa)**: better error message for undefined inputs (pytorch#2540) <Yuxin Wu> - **[8afff0e9](onnx/onnx@8afff0e9)**: bump ORT version (pytorch#2538) <Lu Fang> - **[3d9ca57e](onnx/onnx@3d9ca57e)**: fix name of directory (pytorch#2537) <Prasanth Pulavarthi> - **[df8fa2c9](onnx/onnx@df8fa2c9)**: Repository guidelines (pytorch#2539) <Prasanth Pulavarthi> - **[49cc2f02](onnx/onnx@49cc2f02)**: Update CircleCI job to use Python3.6 (pytorch#2527) <bddppq> - **[25ff79a4](onnx/onnx@25ff79a4)**: Fix wrong model version, it's not 12 (the onnx_opset_version()), not 11 (the opset version of the latest stable), but 10 (pytorch#2478) <daquexian> - **[7cebaed5](onnx/onnx@7cebaed5)**: Fix Windows py3.5 CI (pytorch#2529) <bddppq> - **[eddae00e](onnx/onnx@eddae00e)**: Correct the order of arguments of InferShapes (pytorch#2500) <Shinichiro Hamaji> - **[41b5afe6](onnx/onnx@41b5afe6)**: Include <ostream> in common/status.h (pytorch#2519) <Casey Carter> - **[423f1977](onnx/onnx@423f1977)**: add 8 bit support to maxpool op (pytorch#2510) <Ashwini Khade> - **[78593c2f](onnx/onnx@78593c2f)**: add 8 bit support to reducemin and reducemax ops (pytorch#2516) <Ashwini Khade> Test Plan: cont build Reviewed By: benoitsteiner Differential Revision: D19380034 fbshipit-source-id: ddce8450864a611773b2a32e2f0254c9bb6b6906
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.
Implements double backwards for the following nn modules:
pooling: AdaptiveAvgPool1d, AdaptiveAvgPool2d, FractionalMaxPool2d
unpooling: MaxUnpool1d, MaxUnpool2d
padding: ReplicationPad2d, ReplicationPad3d, ReflectionPad2d.