Skip to content

[Bug Report] Flash decode with GQA on Llama 3.1 8B shape does not work with (8,7) grid size #5798

[Bug Report] Flash decode with GQA on Llama 3.1 8B shape does not work with (8,7) grid size

[Bug Report] Flash decode with GQA on Llama 3.1 8B shape does not work with (8,7) grid size #5798

Annotations

1 warning

gh-slack-bridge

succeeded Dec 8, 2024 in 3s