implementation detail · filed under software implementation
FLASHMLA_SPARSE
A named sparse attention backend selected through the attention-backend option.
Also called FLASHMLA_SPARSE attention backend.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
--attention-backend FLASHMLA_SPARSE
usedsoftware implementationin Hy4-previewvLLM
Filed alongside
Other methods under software implementation :: kernel & quantization library.