Can I optimize manba—especially Selective Scan—to further accelerate inference?
Can I optimize manba—especially Selective Scan—to further accelerate inference?