-
Notifications
You must be signed in to change notification settings - Fork 21.9k
Issues: pytorch/pytorch
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.
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
Author
Label
Projects
Milestones
Assignee
Sort
Issues list
[MPS] Bugs related to incorrect/lacking error checking
module: mps
Related to Apple Metal Performance Shaders framework
F.nll_loss
errors with empty tensor
module: empty tensor
module: error checking
#134431
opened Aug 26, 2024 by
hvaara
[MPS] issue that returns an incorrect result silently
module: mps
Related to Apple Metal Performance Shaders framework
F.interpolate
with mode="nearest-exact"
differs from expected output
module: correctness (silent)
#134430
opened Aug 25, 2024 by
hvaara
[MPS] Empty Bugs related to incorrect/lacking error checking
module: mps
Related to Apple Metal Performance Shaders framework
float32
input to nn.BatchNorm2d
mistaken as float64
module: empty tensor
module: error checking
#134423
opened Aug 25, 2024 by
hvaara
PYTORCH_ENABLE_MPS_FALLBACK
does not appear to work for nn.Conv1d
module: mps
#134416
opened Aug 25, 2024 by
a2aaron
[MPS] MPSNDArray error: product of dimension sizes > 2**32
module: mps
Related to Apple Metal Performance Shaders framework
module: 64-bit
Problems related to incorrectly using 32-bit integers when 64-bit is needed (e.g., 8G tensors)
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#134177
opened Aug 21, 2024 by
Tim-Roy
Binary Cross Entropy crashes kernel on apple silicon with singleton dimensions
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
#133625
opened Aug 15, 2024 by
Derekmod
[MPS] Incorrect result from batch norm with sliced inputs
high priority
module: correctness (silent)
issue that returns an incorrect result silently
module: mps
Related to Apple Metal Performance Shaders framework
module: regression
It used to work, and now it doesn't
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
convolution_overrideable not implemented when executing VAEEncode in ComfyUI
module: PrivateUse1
private use
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#133183
opened Aug 11, 2024 by
spaetische
Leaky ReLU of empty tensor doesn't work on MPS
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
#133179
opened Aug 11, 2024 by
JohnGoldenGardiner
MPS clamp_min on older MacOS
module: correctness (silent)
issue that returns an incorrect result silently
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
#133100
opened Aug 9, 2024 by
wgifford
Silent error in torch.Tensor.sum on mps
module: correctness (silent)
issue that returns an incorrect result silently
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
#132732
opened Aug 6, 2024 by
nisheethlahoti
Pytorch aten::im2col not currently supported on the MPS backend
enhancement
Not as big of a feature, but technically not a bug. Should be easy to fix
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
#132711
opened Aug 5, 2024 by
plops655
Add support for float8 dtypes for the MPS backend
feature
A request for a proper, new feature.
module: float8
For torch.float8_e5m2 and torch.float8_e4m3
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
#132624
opened Aug 5, 2024 by
lisiyizu
MPS lgamma function changes results when using broadcasting
module: correctness (silent)
issue that returns an incorrect result silently
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
#132605
opened Aug 4, 2024 by
canergen
MPS backend leaks memory when input sizes vary
enhancement
Not as big of a feature, but technically not a bug. Should be easy to fix
module: memory usage
PyTorch is using more memory than it should, or it is leaking memory
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
#132596
opened Aug 4, 2024 by
llllvvuu
[MPS] Memory leak in PyTorch is using more memory than it should, or it is leaking memory
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
nn.Linear
high priority
module: memory usage
#132332
opened Jul 31, 2024 by
hvaara
torch.nn.functional.interpolate completely broken with torch.jit.script and torch.fx
oncall: jit
Add this issue/PR to JIT oncall triage queue
#132240
opened Jul 31, 2024 by
domef
MPS backend breaking on llama 3.1 8B on Macbook M3
high priority
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
#131865
opened Jul 26, 2024 by
yliapis
torch.layer_norm()
gives wrong results on MPS if applied on a slice of tensor
module: correctness (silent)
#131750
opened Jul 25, 2024 by
henrysky
MPS gives incorrect result when torch.nn.functional.softplus follows moveaxis
module: correctness (silent)
issue that returns an incorrect result silently
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
#131736
opened Jul 25, 2024 by
JohnGoldenGardiner
MPS ConvTranspose 3D Support
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
#130256
opened Jul 8, 2024 by
valosekj
Assertion failed: (0 <= mpsAxis && mpsAxis < 4 && "Runtime canonicalization must simplify reduction axes to minor 4 dimensions."), function getKernelAxes, file GPUReductionOps.mm, line 31.
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
#129960
opened Jul 2, 2024 by
dbl001
MPS Sparse Support
enhancement
Not as big of a feature, but technically not a bug. Should be easy to fix
module: mps
Related to Apple Metal Performance Shaders framework
module: sparse
Related to torch.sparse
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#129842
opened Jun 30, 2024 by
rootjalex
torch.onnx.export - Related to torch.onnx
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
repeat_interleave
produces invalid model
module: onnx
#128505
opened Jun 12, 2024 by
xenova
[RFC] A device-agnostic Python runtime API design for stream-based accelerators
intel
This tag is for PR from Intel
module: python frontend
For issues relating to PyTorch's Python frontend
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#128403
opened Jun 11, 2024 by
guangyey
Previous Next
ProTip!
Adding no:label will show everything without a label.