How to modify the resolution of generated video?

[sheldonn2]

Very interesting job! When I used it, I found that the model could only generate videos with a resolution of 256. How should I modify the parameters to make the model generate videos with a higher resolution?

I tried changing the "resolution" parameter in the config, but it didn't work.

Looking forward to your reply, thank you!

[agusgun]

I don't think the pretrained model supports arbitrary resolution.
You can try to change it manually by setting cfg.resolution=<YOUR_RESOLUTION> and cfg.misc_size=<YOUR_RESOLUTION>.
But, if you see the model in unet_sd.py line 1230 for example, you need to change the AdaptiveAvgPool layer to use your specific resolution also. It may affect the performance.