Search results

BatchNorm1d exhibits NaN outputs on GPU due to extreme input values causing numerical instability.

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: nnRelated to torch.nnRelated to torch.nn

topic: fuzzer

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#162489 In pytorch/pytorch;

· qiqicliff opened on Sep 9, 2025

crash in torch.nn.functional.interpolate

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: error checkingBugs related to incorrect/lacking error checkingBugs related to incorrect/lacking error checking

module: nnRelated to torch.nnRelated to torch.nn

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#162335 In pytorch/pytorch;

· cx104906 opened on Sep 6, 2025

Floating point exception in torch.nn.PixelShuffle

module: complexRelated to complex number support in PyTorchRelated to complex number support in PyTorch

module: crashProblem manifests as a hard crash, as opposed to a RuntimeErrorProblem manifests as a hard crash, as opposed to a RuntimeError

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: nnRelated to torch.nnRelated to torch.nn

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#162251 In pytorch/pytorch;

· cx104906 opened on Sep 5, 2025

Inconsistent handling of <code>-0.0</code> between CPU and CUDA for many operators

module: cudaRelated to torch.cuda, and CUDA support in generalRelated to torch.cuda, and CUDA support in general

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: NaNs and InfsProblems related to NaN and Inf handling in floating pointProblems related to NaN and Inf handling in floating point

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#162235 In pytorch/pytorch;

· rookieLiu2018 opened on Sep 5, 2025

Segmentation fault in torch.nn.LazyConv1d

module: crashProblem manifests as a hard crash, as opposed to a RuntimeErrorProblem manifests as a hard crash, as opposed to a RuntimeError

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: error checkingBugs related to incorrect/lacking error checkingBugs related to incorrect/lacking error checking

topic: fuzzer

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#161875 In pytorch/pytorch;

· qiqicliff opened on Sep 1, 2025

Floating point exception in torch.Tensor.mvlgamma_

module: crashProblem manifests as a hard crash, as opposed to a RuntimeErrorProblem manifests as a hard crash, as opposed to a RuntimeError

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: error checkingBugs related to incorrect/lacking error checkingBugs related to incorrect/lacking error checking

module: intelSpecific to x86 architectureSpecific to x86 architecture

module: specialFunctions with no exact solutions, analogous to those in scipy.specialFunctions with no exact solutions, analogous to those in scipy.special

topic: fuzzer

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#161871 In pytorch/pytorch;

· cx104906 opened on Sep 1, 2025

torch.nn.FractionalMaxPool3d occurs bug with big kernel_size

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: error checkingBugs related to incorrect/lacking error checkingBugs related to incorrect/lacking error checking

topic: fuzzer

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#161853 In pytorch/pytorch;

· qiqicliff opened on Aug 31, 2025

Invalid __shared__ read of size 8 bytes in torch.histc

module: crashProblem manifests as a hard crash, as opposed to a RuntimeErrorProblem manifests as a hard crash, as opposed to a RuntimeError

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: sanitizers

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#161760 In pytorch/pytorch;

· cx104906 opened on Aug 29, 2025

boolean index_add: Discrepancy between CPU and MPS backend

low priorityWe're unlikely to get around to doing this in the near futureWe're unlikely to get around to doing this in the near future

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: mpsRelated to Apple Metal Performance Shaders frameworkRelated to Apple Metal Performance Shaders framework

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#161524 In pytorch/pytorch;

· manuelcandales opened on Aug 26, 2025

Discrepancy between PyTorch and TensorRT when casting from FP32 to BF16 causes significant accuracy mismatch

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: onnxRelated to torch.onnxRelated to torch.onnx

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#160856 In pytorch/pytorch;

· KarlDe1 opened on Aug 18, 2025

MPS copy_ produces incorrect output when input has more than one element

module: correctness (silent)issue that returns an incorrect result silentlyissue that returns an incorrect result silently

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: mpsRelated to Apple Metal Performance Shaders frameworkRelated to Apple Metal Performance Shaders framework

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

Bug

#160744 In pytorch/pytorch;

· manuelcandales opened on Aug 15, 2025

pad_sequence overflows when padding with larger value than int64.max - 512

module: edge casesAdversarial inputs unlikely to occur in practiceAdversarial inputs unlikely to occur in practice

module: python frontendFor issues relating to PyTorch's Python frontendFor issues relating to PyTorch's Python frontend

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#159363 In pytorch/pytorch;

· skaae opened on Jul 29, 2025