Make Conv{1,2,3}dOptions and ConvTranspose{1,2,3}dOptions different classes #31005
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
Contributor
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@yf225 has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
Contributor
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@yf225 has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
pbelevich
approved these changes
Dec 11, 2019
yf225
pushed a commit
that referenced
this pull request
Dec 12, 2019
…lasses (#31005) Summary: Currently, both `Conv{1,2,3}dOptions` and `ConvTranspose{1,2,3}dOptions` are aliases of the `ConvOptions<{1,2,3}>` class, which causes confusion because the `ConvOptions` class has parameters such as `transposed` that shouldn't be exposed to the end user. (This has caused issues such as #30931.) This PR makes the following improvements: 1. Rename the original `torch::nn::ConvOptions<N>` class to `torch::nn::detail::ConvNdOptions<N>` class, to signify that it's an implementation detail and should not be used publicly. 2. Create new classes `torch::nn::ConvOptions<N>` and `torch::nn::ConvTransposeOptions<N>`, which have parameters that exactly match the constructor of `torch.nn.Conv{1,2,3}d` and `torch.nn.ConvTranspose{1,2,3}d` in Python API. Pull Request resolved: #31005 Differential Revision: D18898048 Pulled By: yf225 fbshipit-source-id: 7663d646304c8cb004ca7f4aa4e70d3612c7bc75
yf225
pushed a commit
that referenced
this pull request
Dec 12, 2019
…lasses (#31005) Summary: Currently, both `Conv{1,2,3}dOptions` and `ConvTranspose{1,2,3}dOptions` are aliases of the `ConvOptions<{1,2,3}>` class, which causes confusion because the `ConvOptions` class has parameters such as `transposed` that shouldn't be exposed to the end user. (This has caused issues such as #30931.) This PR makes the following improvements: 1. Rename the original `torch::nn::ConvOptions<N>` class to `torch::nn::detail::ConvNdOptions<N>` class, to signify that it's an implementation detail and should not be used publicly. 2. Create new classes `torch::nn::ConvOptions<N>` and `torch::nn::ConvTransposeOptions<N>`, which have parameters that exactly match the constructor of `torch.nn.Conv{1,2,3}d` and `torch.nn.ConvTranspose{1,2,3}d` in Python API. Pull Request resolved: #31005 Differential Revision: D18898048 Pulled By: yf225 fbshipit-source-id: 7663d646304c8cb004ca7f4aa4e70d3612c7bc75
Contributor
This was referenced Jan 16, 2020
wuhuikx
pushed a commit
to wuhuikx/pytorch
that referenced
this pull request
Jan 30, 2020
…lasses (pytorch#31005) Summary: Currently, both `Conv{1,2,3}dOptions` and `ConvTranspose{1,2,3}dOptions` are aliases of the `ConvOptions<{1,2,3}>` class, which causes confusion because the `ConvOptions` class has parameters such as `transposed` that shouldn't be exposed to the end user. (This has caused issues such as pytorch#30931.) This PR makes the following improvements: 1. Rename the original `torch::nn::ConvOptions<N>` class to `torch::nn::detail::ConvNdOptions<N>` class, to signify that it's an implementation detail and should not be used publicly. 2. Create new classes `torch::nn::ConvOptions<N>` and `torch::nn::ConvTransposeOptions<N>`, which have parameters that exactly match the constructor of `torch.nn.Conv{1,2,3}d` and `torch.nn.ConvTranspose{1,2,3}d` in Python API. Pull Request resolved: pytorch#31005 Differential Revision: D18898048 Pulled By: yf225 fbshipit-source-id: 7663d646304c8cb004ca7f4aa4e70d3612c7bc75
wuhuikx
pushed a commit
to wuhuikx/pytorch
that referenced
this pull request
Jan 30, 2020
…ow to use them in a Sequential module (pytorch#32223) Summary: Following changes in pytorch#31005. Pull Request resolved: pytorch#32223 Differential Revision: D19415328 Pulled By: yf225 fbshipit-source-id: f6f74f10ba3b5cc7e1a92f8b02ea4c9747018ae8
ttumiel
pushed a commit
to ttumiel/pytorch
that referenced
this pull request
Mar 4, 2020
…ow to use them in a Sequential module (pytorch#32223) Summary: Following changes in pytorch#31005. Pull Request resolved: pytorch#32223 Differential Revision: D19415328 Pulled By: yf225 fbshipit-source-id: f6f74f10ba3b5cc7e1a92f8b02ea4c9747018ae8
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.
Currently, both
Conv{1,2,3}dOptionsandConvTranspose{1,2,3}dOptionsare aliases of theConvOptions<{1,2,3}>class, which causes confusion because theConvOptionsclass has parameters such astransposedthat shouldn't be exposed to the end user. (This has caused issues such as #30931.) This PR makes the following improvements:torch::nn::ConvOptions<N>class totorch::nn::detail::ConvNdOptions<N>class, to signify that it's an implementation detail and should not be used publicly.torch::nn::ConvOptions<N>andtorch::nn::ConvTransposeOptions<N>, which have parameters that exactly match the constructor oftorch.nn.Conv{1,2,3}dandtorch.nn.ConvTranspose{1,2,3}din Python API.This PR is BC-breaking in the following way:
Conv{1,2,3}dOptionsno longer has thetransposedargument. Users should migrate their code to useConvTranspose{1,2,3}dlayers instead, if they havetransposedoriginally set totrueinConv{1,2,3}dOptions. Note thatConvTranspose{1,2,3}dcannot be used in aSequentialmodule becauseSequentialmodule doesn't support modules with forward method that can take optional arguments. Users should create their own wrapper forConvTranspose{1,2,3}dand have its forward method just accept a tensor, if they want to use it in aSequentialmodule.For example: