Search results

at::Tensor::to_sparse_csr() does not produce correct result in Debug configuration when tensor is big sparse CSC on CUDA device in LibTorch 2.8.0

module: cppRelated to C++ APIRelated to C++ API

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

module: debug-buildRelated to building and testing PyTorch in debug modeRelated to building and testing PyTorch in debug mode

module: sparseRelated to torch.sparseRelated to torch.sparse

module: windowsWindows support for PyTorchWindows support for PyTorch

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.

#163502 In pytorch/pytorch;

· marek-racev opened on Sep 22, 2025

Add Flash Dynamic Mask Attention (FDMA) as an SDPA backend for sparse, extremely long-context attention

module: sdpaAll things related to torch.nn.functional.scaled_dot_product_attentiionAll things related to torch.nn.functional.scaled_dot_product_attentiion

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#163400 In pytorch/pytorch;

· LoserCheems opened on Sep 20, 2025

NotImplementedError: "addmm_sparse_cuda" not implemented for 'Half'

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

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#162674 In pytorch/pytorch;

· wangmiaowei opened on Sep 11, 2025

[sparse] <code>torch._sparse_broadcast_to</code> creates false-positive coalesced tensors with inputs of shape <code>()</code>

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#162482 In pytorch/pytorch;

· nikitaved opened on Sep 9, 2025

<code>reshape</code> support for sparse tensors (needed for <code>einsum</code>)

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: sparseRelated to torch.sparseRelated to torch.sparse

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.

#161357 In pytorch/pytorch;

· Giuseppe499 opened on Aug 23, 2025

Support optimisation of sparse parameters via LBFGS

module: optimizerRelated to torch.optimRelated to torch.optim

module: sparseRelated to torch.sparseRelated to torch.sparse

needs designWe want to add this feature but we need to figure out how firstWe want to add this feature but we need to figure out how first

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.

#160637 In pytorch/pytorch;

· Thomas-Christie opened on Aug 14, 2025

torch.quantize_per_tensor not implemented for SparseCPU tensors

module: sparseRelated to torch.sparseRelated to torch.sparse

oncall: quantizationQuantization support in PyTorchQuantization support in PyTorch

Status: Open.

#160631 In pytorch/pytorch;

· springcrane opened on Aug 14, 2025

Enhance output layouts of all-sparse matrix multiplications

matrix multiplication

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#160517 In pytorch/pytorch;

· Aidyn-A opened on Aug 13, 2025

Different behavior between sparse and dense tensors with broadcasting multiplication.

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#158861 In pytorch/pytorch;

· roccomoretti opened on Jul 22, 2025

sparse_csr bfloat16 matrix multiplication backward is 10x slower than float16

module: bfloat16

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

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#157808 In pytorch/pytorch;

· zbh2047 opened on Jul 8, 2025

Issue when Using SparseMPS backend (especially the operation aten::_sparse_coo_tensor_with_dims_and_tensors) for Whisper model

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

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#156540 In pytorch/pytorch;

· Stan370 opened on Jun 21, 2025

[A100][cusparselt] Non-determinstic correctness issue with some algorithms

module: sparseRelated to torch.sparseRelated to torch.sparse

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.

#155333 In pytorch/pytorch;

· danthe3rd opened on Jun 6, 2025