FlashMLA support

[EricLBuehler]

Support FlashMLA for improved throughput for MLA models (DeepSeek V2, V3/R1) on CUDA.

EricLBuehler/candle#74

https://github.com/deepseek-ai/FlashMLA

[github-actions]

Code Metrics Report

  ===============================================================================
 Language            Files        Lines         Code     Comments       Blanks
===============================================================================
 C Header                2           35           28            0            7
 Dockerfile              1           41           22           10            9
 JSON                   12          105          104            0            1
 Python                 69         2926         2534           77          315
 Shell                   1           58           22           18           18
 Plain Text              3         3723            0         2413         1310
 TOML                   19          527          489            2           36
 YAML                    2           21           19            2            0
-------------------------------------------------------------------------------
 Jupyter Notebooks       4            0            0            0            0
 |- Markdown             2           77           32           31           14
 |- Python               2          205          178            1           26
 (Total)                            282          210           32           40
-------------------------------------------------------------------------------
 Markdown               48         3894            0         2953          941
 |- BASH                 6          103          100            0            3
 |- JSON                 1           12           12            0            0
 |- Python               7          121          109            0           12
 |- Rust                15          512          433            0           79
 |- TOML                 2           75           63            0           12
 (Total)                           4717          717         2953         1047
-------------------------------------------------------------------------------
 Rust                  318       103707        92870         2075         8762
 |- Markdown           156         1763           25         1604          134
 (Total)                         105470        92895         3679         8896
===============================================================================
 Total                 479       115037        96088         7550        11399
===============================================================================