8000 MPS ConvTranspose 3D Support · Issue #130256 · pytorch/pytorch · GitHub
[go: up one dir, main page]
Skip to content

MPS ConvTranspose 3D Support #130256

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

Jump to bottom
Open
valosekj opened this issue Jul 8, 2024 · 1 com 8000 ment · May be fixed by #145366
Open

MPS ConvTranspose 3D Support #130256

valosekj opened this issue Jul 8, 2024 · 1 comment · May be fixed by #145366
Labels
enhancement Not as big of a feature, but technically not a bug. Should be easy to fix module: convolution Problems related to convolutions (THNN, THCUNN, CuDNN) module: mps Related to Apple Metal Performance Shaders framework triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@valosekj
Copy link
valosekj commented Jul 8, 2024 •
edited by pytorch-bot bot
Loading

🚀 The feature, motivation and pitch

torch.nn.functional.conv_transpose3d is not currently supported on MPS (Apple Silicon):

import torch
import torch.nn.functional as F

# Create a random input tensor on MPS
device = torch.device("mps")
input_tensor = torch.randn(1, 1, 8, 8, 8, device=device)

# Define a random weight tensor
weight = torch.randn(1, 1, 3, 3, 3, device=device)

# Perform the 3D transposed convolution
output = F.conv_transpose3d(input_tensor, weight)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
RuntimeError: ConvTranspose 3D is not supported on MPS
>>> print(torch.__version__)
2.3.1

This limits, for example, the usage of medical image segmentation framework nnUNet to torch.device("cpu").

Alternatives

No response

Additional context

No response

cc @kulinseth @albanD @malfet @DenisVieriu97 @jhavukainen
@valosekj valosekj mentioned this issue Jul 8, 2024
Open
@zou3519 zou3519 added module: convolution Problems related to convolutions (THNN, THCUNN, CuDNN) triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module module: mps Related to Apple Metal Performance Shaders framework labels Jul 9, 2024
@valosekj valosekj mentioned this issue Aug 8, 2024
Closed
@Twenty3hree
Copy link

Hows about it now?

@mlaves mlaves linked a pull request Jan 22, 2025 that will close this issue
Open
@jhavukainen jhavukainen added the enhancement Not as big of a feature, but technically not a bug. Should be easy to fix label Jan 24, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement Not as big of a feature, but technically not a bug. Should be easy to fix module: convolution Problems related to convolutions (THNN, THCUNN, CuDNN) module: mps Related to Apple Metal Performance Shaders framework triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

removed check for ConvTranspose3D on MPS mlaves/pytorch
4 participants
@zou3519 @Twenty3hree @valosekj @jhavukainen
0