perf(Gemma4): reduce comms in sliding window, FA implementation, (#2984)
* Gemma4: reduce comms in sliding window, FA implementation, Signed-off-by: Nathan Azrak <nathan.azrak@gmail.com> * testing suite Signed-off-by: Nathan Azrak <nathan.azrak@gmail.com> * fix(gemma4): preserve vision CP attention semantics Signed-off-by: HuiyingLi <willwin.lee@gmail.com> --------- Signed-off-by: Nathan Azrak <nathan.azrak@gmail.com> Signed-off-by: HuiyingLi <willwin.lee@gmail.com> Co-authored-by: Alexandros Koumparoulis <153118171+akoumpa@users.noreply.github.com> Co-authored-by: HuiyingLi <willwin.lee@gmail.com>
N
Nathan Azrak committed
c63a6633f48fd68a4ff738c63ce05389446533ec
Parent: da5eba2
Committed by GitHub <noreply@github.com>
on 8/20/2026, 5:12:44 PM