You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
首先感谢团队做出的卓越贡献,controlnet应用速度大幅度提升了。但是目前貌似teacache等加速技术就失效了,使用时没看到明显的效果。
另外分区绘制的时候如图,单纯只是接入Nunchaku FLUX DiT Loader这个节点作为开始都会报错,这个分区绘制速度特别慢,特别需要加速:
The size of tensor a (266) must match the size of tensor b (256) at non-singleton dimension 1