[Kernels] Refactor flash attention for continuous batching. - Githubissues

intel / xFasterTransformer

Apache License 2.0

344 stars 60 forks source link

[Kernels] Refactor flash attention for continuous batching. #361

Closed abenmao closed 4 months ago