search

MCG-NJU / MGMAE

[ICCV 2023] MGMAE: Motion Guided Masking for Video Masked Autoencoding
MIT License
20 stars 0 forks source link

How to obtain the reconstructed image for inference and masked #2

Open hzxie99 opened 10 months ago

hzxie99 commented 10 months ago

Hello, I very much agree with your work. I would like to know how to obtain the schematic diagram of the reconstructed image and the mask image during the experiment. Because I'm just getting started. I really appreciate it if you can help me with this question.

congee524 commented 9 months ago

This part of the code I didn't organize. I did write a special train_one_epoch function that saves the input image, mask map, and reconstruction results of the process at the first iteration, and then just exits.