Validating Pytorch Custom Ops

Ensures new Attention Gym eager, Triton, CuTeDSL, and external-library implementations are torch.compile-friendly and correctly registered. Use when adding or reviewing an implementation or custom kernel; covers custom_op, triton_op, fake tensors, autograd, opcheck, full-graph compilation, dynamic shapes, and CUDA Graphs.

meta-pytorch Updated
Two new Flexys—Sparse Attention and Linear Attention—enter Attention Gym while the original Flexys train

File contents

meta-pytorch/attention-gym commit 6614a43ef4

Frequently asked questions

npx get-skill add meta-pytorch/validating-pytorch-custom-ops