Search results

RuntimeError: ptrDesc->finalize() with conv1d

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

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

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.

#161778 In pytorch/pytorch;

· rangehow opened on Aug 29, 2025

[Track] Optimizing Groupwise convolution for aarch64

module: armRelated to ARM architectures builds of PyTorch. Includes Apple M1Related to ARM architectures builds of PyTorch. Includes Apple M1

module: bfloat16

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

module: performanceIssues related to performance, either of kernel code or framework glueIssues related to performance, either of kernel code or framework glue

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.

#158923 In pytorch/pytorch;

· aditew01 opened on Jul 23, 2025

Inefficient 2D convolution compared to JAX

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

module: nnRelated to torch.nnRelated to torch.nn

module: performanceIssues related to performance, either of kernel code or framework glueIssues related to performance, either of kernel code or framework glue

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.

#157334 In pytorch/pytorch;

· null-pointer-access opened on Jul 1, 2025

3D Conv Slow with Large Tensors (> 2**31) Elements

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

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

module: performanceIssues related to performance, either of kernel code or framework glueIssues related to performance, either of kernel code or framework glue

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.

#155218 In pytorch/pytorch;

· viengiaan opened on Jun 5, 2025

integer convolution

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

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

needs researchWe need to decide whether or not this merits inclusion, based on research worldWe need to decide whether or not this merits inclusion, based on research world

oncall: quantizationQuantization support in PyTorchQuantization support in PyTorch

Status: Open.

#154354 In pytorch/pytorch;

· epii2zero opened on May 26, 2025

<code>torch.nn.functional.conv_transpose2d</code> has inconsistent handling of <code>float16</code> overflow on CPU

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

module: halfRelated to float16 half-precision floatsRelated to float16 half-precision floats

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.

#153700 In pytorch/pytorch;

· jiren-the-gray opened on May 16, 2025

Floating Point exception in Convolution with disabled SMT

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

module: dependency bugProblem is not caused by us, but caused by an upstream library we useProblem is not caused by us, but caused by an upstream library we use

module: nnRelated to torch.nnRelated to torch.nn

module: nnpackRelated to our NNPack integrationRelated to our NNPack integration

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.

#153139 In pytorch/pytorch;

· Flamefire opened on May 8, 2025

conv2d with int8 on CUDA: GET was unable to find an engine to execute this computation

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

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

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.

#152992 In pytorch/pytorch;

· c-f-h opened on May 6, 2025

conv2d fp8 support

enhancementNot as big of a feature, but technically not a bug. Should be easy to fixNot as big of a feature, but technically not a bug. Should be easy to fix

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

module: floatx (formerly float8)For torch.float8_e5m2 and torch.float8_e4m3 and other sub 8-bit float typesFor torch.float8_e5m2 and torch.float8_e4m3 and other sub 8-bit float types

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.

#150614 In pytorch/pytorch;

· sipie800 opened on Apr 3, 2025

NotImplementedError: Output channels &gt; 65536 not supported at the MPS device.

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

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.

#144445 In pytorch/pytorch;

· AimoneAndex opened on Jan 9, 2025

<code>MemoryDenyWriteExecute</code> in systemd service causes <code>RuntimeError: could not create a primitive</code>

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

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

module: mkldnnRelated to Intel IDEEP or oneDNN (a.k.a. mkldnn) integrationRelated to Intel IDEEP or oneDNN (a.k.a. mkldnn) integration

needs reproductionEnsure you have actionable steps to reproduce the issue. Someone else needs to confirm the repro.Ensure you have actionable steps to reproduce the issue. Someone else needs to confirm the repro.

security

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.

#143651 In pytorch/pytorch;

· MatthewCroughan opened on Dec 20, 2024

Segmentation fault (core dumped) in <code>conv3d</code>

module: convolutionProblems related to convolutions (THNN, THCUNN, CuDNN)Problems related to convolutions (THNN, THCUNN, CuDNN)

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

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

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.

#143481 In pytorch/pytorch;

· LongZE666 opened on Dec 18, 2024