Search results

segment_coo / segment_csr do not reject out-of-range index metadata before CUDA execution

Status: Open.

#511 In rusty1s/pytorch_scatter;

· supermarkli opened on Jun 22, 2026

New release

Status: Open.

#509 In rusty1s/pytorch_scatter;

· Andrew-S-Rosen opened on May 31, 2026

Add Ascend NPU support to pytorch_scatter

Status: Open.

#508 In rusty1s/pytorch_scatter;

· BigFishQin opened on May 6, 2026

Building torch_scatter from source leads to "Detected that PyTorch and torch_scatter were compiled with different CUDA versions."

Status: Open.

#503 In rusty1s/pytorch_scatter;

· JCBrouwer opened on Mar 3, 2026

Requirements-adding support for torch2.9+cu130

Status: Open.

#500 In rusty1s/pytorch_scatter;

· irrikrlla opened on Oct 28, 2025

Non-blocking gather_csr_cuda()

enhancement

Status: Open.

#305 In rusty1s/pytorch_scatter;

· rBenke opened on Jun 23, 2022

Empty second-order derivative (= Hessian) for the segment_* reductions

bug

Status: Open.

#299 In rusty1s/pytorch_scatter;

· jeanfeydy opened on May 19, 2022

Possible bug in half-precision scatter

bug

Status: Open.

#269 In rusty1s/pytorch_scatter;

· noahstier opened on Feb 13, 2022

Issues from importing torch in setup.py

enhancement

Status: Open.

#265 In rusty1s/pytorch_scatter;

· cthoyt opened on Feb 2, 2022

PyTorch1.10+cu113 CUDA Graph for torch_scatter raises RuntimeError w/o argument out

enhancement

Status: Open.

#244 In rusty1s/pytorch_scatter;

· liulixinkerry opened on Oct 23, 2021