[C++ API parity] Multi-Label Soft Margin loss #27669
Conversation
CarMiranda
requested review from
ebetica,
goldsborough and
yf225
as code owners
yf225
left a comment
yf225
left a comment
There was a problem hiding this comment.
Thanks a lot for the fantastic work @CarMiranda! I left some very minor comments.
| // ============================================================================ | ||
|
|
||
| MultiLabelSoftMarginLossImpl::MultiLabelSoftMarginLossImpl( | ||
| const torch::nn::MultiLabelSoftMarginLossOptions& options_) |
There was a problem hiding this comment.
clang-tidy seems to complain about this line, and we can do the following to get around it:
| const torch::nn::MultiLabelSoftMarginLossOptions& options_) | |
| const torch::nn::MultiLabelSoftMarginLossOptions& options_) // NOLINT(modernize-pass-by-value) |
|
@pytorchbot rebase this please |
| /// Creates a criterion that optimizes a multi-label one-versus-all | ||
| /// loss based on max-entropy, between input :math:`x` and target :math:`y` of size | ||
| /// :math:`(N, C)`. | ||
| struct TORCH_API MultiLabelSoftMarginLossImpl : Module { |
There was a problem hiding this comment.
My sincere apologies for missing this earlier: for all torch::nn layers we will need to subclass it from Cloneable in the following way:
| struct TORCH_API MultiLabelSoftMarginLossImpl : Module { | |
| struct TORCH_API MultiLabelSoftMarginLossImpl : public torch::nn::Cloneable<MultiLabelSoftMarginLossImpl> { |
otherwise module->clone() won't work on them. I think for all currently open PRs we should change this, and I will fix the modules that are already in the codebase (e.g. L1Loss).
There was a problem hiding this comment.
Okay, no problem! Thanks for the review!
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.
Summary: In accordance with pytorch#25883, I added the `MultiLabelSoftMarginLoss` module and `multilabel_soft_margin_loss` functional. It looks like there isn't a C++ ATen implementation of `multilabel_soft_margin_loss`, so I translated the python version, which does not rely on a C/C++ backend either. Pull Request resolved: pytorch#27669 Differential Revision: D17907608 Pulled By: yf225 fbshipit-source-id: ccb02951e009973c2adbe604593ce929f10c39eb
5 participants
In accordance with #25883, I added the
MultiLabelSoftMarginLossmodule andmultilabel_soft_margin_lossfunctional.It looks like there isn't a C++ ATen implementation of
multilabel_soft_margin_loss, so I translated the python version, which does not rely on a C/C++ backend either.