Skip to content

Deprecate old fft functions #44876

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
peterbell10 wants to merge 15 commits into from
Closed

Deprecate old fft functions #44876

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
363d6a7
Deprecate old fft functions
peterbell10 Sep 17, 2020
12ed8df
Update on "Deprecate old fft functions"
peterbell10 Sep 17, 2020
18f360d
Update on "Deprecate old fft functions"
peterbell10 Sep 17, 2020
a17490d
Update on "Deprecate old fft functions"
peterbell10 Sep 19, 2020
6a57fea
Update on "Deprecate old fft functions"
peterbell10 Sep 19, 2020
953f193
Update on "Deprecate old fft functions"
peterbell10 Sep 19, 2020
8b271dd
Update on "Deprecate old fft functions"
peterbell10 Sep 19, 2020
94bac81
Update on "Deprecate old fft functions"
peterbell10 Sep 22, 2020
db2c1d2
Update on "Deprecate old fft functions"
peterbell10 Sep 22, 2020
4d62eb3
Update on "Deprecate old fft functions"
peterbell10 Sep 22, 2020
b1ab343
Update on "Deprecate old fft functions"
peterbell10 Sep 23, 2020
cc898c1
Update on "Deprecate old fft functions"
peterbell10 Sep 23, 2020
cde5fbc
Update on "Deprecate old fft functions"
peterbell10 Sep 23, 2020
edaf6d6
Update on "Deprecate old fft functions"
peterbell10 Sep 23, 2020
cd37267
Update on "Deprecate old fft functions"
peterbell10 Sep 23, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 16 additions & 0 deletions aten/src/ATen/native/SpectralOps.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -561,13 +561,21 @@ void _cufft_clear_plan_cache(int64_t device_index) {
}

Tensor fft(const Tensor& self, const int64_t signal_ndim, const bool normalized) {
TORCH_WARN_ONCE(
"The function torch.fft is deprecated and will be removed in PyTorch 1.8. "
"Use the new torch.fft module functions, instead, by importing torch.fft "
"and calling torch.fft.fft or torch.fft.fftn.");
return _fft(self, signal_ndim, /* complex_input */ true,
/* complex_output */ true, /* inverse */ false, {},
normalized ? fft_norm_mode::by_root_n : fft_norm_mode::none,
/* onesided */ false);
}

Tensor ifft(const Tensor& self, const int64_t signal_ndim, const bool normalized) {
TORCH_WARN_ONCE(
"The function torch.ifft is deprecated and will be removed in a future "
"PyTorch release. Use the new torch.fft module functions, instead, by "
"importing torch.fft and calling torch.fft.ifft or torch.fft.ifftn.");
return _fft(self, signal_ndim, /* complex_input */ true,
/* complex_output */ true, /* inverse */ true, {},
normalized ? fft_norm_mode::by_root_n : fft_norm_mode::by_n,
Expand All @@ -576,6 +584,10 @@ Tensor ifft(const Tensor& self, const int64_t signal_ndim, const bool normalized

Tensor rfft(const Tensor& self, const int64_t signal_ndim, const bool normalized,
const bool onesided) {
TORCH_WARN_ONCE(
Copy link
Collaborator

@mruberry mruberry Sep 22, 2020

Choose a reason for hiding this comment

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

Deprecating torch.fft() is the priority (since its name conflicts with the module), but deprecating these is OK if the same pattern is followed. I.e.:

""The function torch.rfft is deprecated and will be removed in a future PyTorch release. Use the new torch.fft module functions, instead, by importing torch.fft and calling torch.fft.rfft or torch.fft.rfftn."

And also updating the torch.rfft and torch.irfft docs.

"The function torch.rfft is deprecated and will be removed in a future "
"PyTorch release. Use the new torch.fft module functions, instead, by "
"importing torch.fft and calling torch.fft.fft or torch.fft.rfft.");
return _fft(self, signal_ndim, /* complex_input */ false,
/* complex_output */ true, /* inverse */ false, {},
normalized ? fft_norm_mode::by_root_n : fft_norm_mode::none,
Expand All @@ -584,6 +596,10 @@ Tensor rfft(const Tensor& self, const int64_t signal_ndim, const bool normalized

Tensor irfft(const Tensor& self, const int64_t signal_ndim, const bool normalized,
const bool onesided, IntArrayRef signal_sizes) {
TORCH_WARN_ONCE(
"The function torch.irfft is deprecated and will be removed in a future "
"PyTorch release. Use the new torch.fft module functions, instead, by "
"importing torch.fft and calling torch.fft.ifft or torch.fft.irfft.");
return _fft(self, signal_ndim, /* complex_input */ true,
/* complex_output */ false, /* inverse */ true, signal_sizes,
normalized ? fft_norm_mode::by_root_n : fft_norm_mode::by_n,
Expand Down
2 changes: 2 additions & 0 deletions docs/source/fft.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,8 @@
.. role:: hidden
:class: hidden-section

.. _torch-fft-module:

torch.fft
=========

Expand Down
26 changes: 26 additions & 0 deletions torch/_torch_docs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8514,6 +8514,12 @@ def merge_dicts(*dicts):

The inverse of this function is :func:`~torch.ifft`.

.. deprecated:: 1.7.0
The function :func:`torch.fft` is deprecated and will be removed in
PyTorch 1.8. Use the new :ref:`torch.fft <torch-fft-module>` module
functions, instead, by importing :ref:`torch.fft <torch-fft-module>` and
calling :func:`torch.fft.fft` or :func:`torch.fft.fftn`.

.. note::
For CUDA tensors, an LRU cache is used for cuFFT plans to speed up
repeatedly running FFT methods on tensors of same geometry with same
Expand Down Expand Up @@ -8617,6 +8623,12 @@ def merge_dicts(*dicts):

The inverse of this function is :func:`~torch.fft`.

.. deprecated:: 1.7.0
The function :func:`torch.ifft` is deprecated and will be removed in a
future PyTorch release. Use the new :ref:`torch.fft <torch-fft-module>`
module functions, instead, by importing :ref:`torch.fft <torch-fft-module>`
and calling :func:`torch.fft.ifft` or :func:`torch.fft.ifftn`.

.. note::
For CUDA tensors, an LRU cache is used for cuFFT plans to speed up
repeatedly running FFT methods on tensors of same geometry with same
Expand Down Expand Up @@ -8705,6 +8717,13 @@ def merge_dicts(*dicts):

The inverse of this function is :func:`~torch.irfft`.

.. deprecated:: 1.7.0
The function :func:`torch.rfft` is deprecated and will be removed in a
future PyTorch release. Use the new :ref:`torch.fft <torch-fft-module>`
module functions, instead, by importing :ref:`torch.fft <torch-fft-module>`
and calling :func:`torch.fft.rfft` for one-sided output, or
:func:`torch.fft.fft` for two-sided output.

.. note::
For CUDA tensors, an LRU cache is used for cuFFT plans to speed up
repeatedly running FFT methods on tensors of same geometry with same
Expand Down Expand Up @@ -8777,6 +8796,13 @@ def merge_dicts(*dicts):

The inverse of this function is :func:`~torch.rfft`.

.. deprecated:: 1.7.0
The function :func:`torch.irfft` is deprecated and will be removed in a
future PyTorch release. Use the new :ref:`torch.fft <torch-fft-module>`
module functions, instead, by importing :ref:`torch.fft <torch-fft-module>`
and calling :func:`torch.fft.irfft` for one-sided input, or
:func:`torch.fft.ifft` for two-sided input.

.. warning::
Generally speaking, input to this function should contain values
following conjugate symmetry. Note that even if :attr:`onesided` is
Expand Down