Commit 73d8740
authored
Add SeerAttention and SlimAttention Paper (#135)
* Add slim-attention: transform KV-cache to K cache only
Signed-off-by: sven <[email protected]>
* Add SeerAttention: learnable sparse attention like NSA(deepseek) MoBA
Signed-off-by: sven <[email protected]>
---------
Signed-off-by: sven <[email protected]>1 parent 8a0ae90 commit 73d8740
1 file changed
+305
-303
lines changed
0 commit comments