About SVD Model Inference Code

[jhj7905]

@yxgeee @xinntao @wzhouxiff @eltociear @yeliudev
Hello, Thank you for releasing great work.
I have questions about SVD Model Inference Code.
Why do you multiply np.array([3, 1, 4]) to pose?? Can you tell me it in details?

MotionCtrl/main/inference/motionctrl_cmcm.py

Line 85 in 30953f7

pose[:, :, -1] = pose[:, :, -1] * np.array([3, 1, 4]) * speed

And if I want to train the model with lower resolution(1024x576 -> 512x320), how do i modify the code?
Thank you in advance

[yshen47]

Same question, "Why do you multiply np.array([3, 1, 4]) to pose??"

[sacred178953]

@yxgeee @xinntao @wzhouxiff @eltociear @yeliudev Hello, Thank you for releasing great work. I have questions about SVD Model Inference Code. Why do you multiply np.array([3, 1, 4]) to pose?? Can you tell me it in details?

MotionCtrl/main/inference/motionctrl_cmcm.py

Line 85 in 30953f7

pose[:, :, -1] = pose[:, :, -1] * np.array([3, 1, 4]) * speed

And if I want to train the model with lower resolution(1024x576 -> 512x320), how do i modify the code? Thank you in advance

Hello,do you have the training code for the image-to-video SVD model? It seems the authors didn't provide it, and I couldn't find it.If you have, could you share it?Thanks