FlexAttention: The Flexibility of PyTorch…

Horace He

Aug 7, 2024

17

1

2

Freeing users from the software lottery tyranny of fused attention implementations.

Read →

1 Comment

Hailey Schoelkopf

Aug 7

Would love to see the post on nondeterminism! Think it would be a very useful resource

Expand full comment

Reply

Share

#nojs-banner { position: fixed; bottom: 0; left: 0; padding: 16px 16px 16px 32px; width: 100%; box-sizing: border-box; background: red; color: white; font-family: -apple-system, "Segoe UI", Roboto, Helvetica, Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol"; font-size: 13px; line-height: 13px; } #nojs-banner a { color: inherit; text-decoration: underline; } This site requires JavaScript to run correctly. Please turn on JavaScript or unblock scripts

Thonk From First Principles

FlexAttention: The Flexibility of PyTorch…