xla
Update ragged paged attention kernel to prevent vmem oom
#9346
Merged

Commits
  • calculate actual vmem usage after getting kernel shapes
    Chenyaaang committed 305 days ago
  • Merge remote-tracking branch 'cy-fork/master' into vmem-oom
    Chenyaaang committed 305 days ago
  • calculate actual kerrnel vem usage
    Chenyaaang committed 300 days ago
  • Merge remote-tracking branch 'origin/master' into vmem-oom
    Chenyaaang committed 297 days ago
Loading