For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).
Mojo module
mla_sparse_attention
Runs prepared sparse attention with the SM100 decode attention core.
Functionsโ
- โ
mla_sparse_attention: Runs attention over packed physical index rows and per-query lengths. - โ
mla_sparse_attention_flat: Runs independent physical sparse rows with the shared attention core.