Regarding the issue of inferring multiple images, thank you for your excellent work

[lxin98]

What are the differences between train_bit.py and inference.py? Can it be understood that inference.py handles three continuous images? Is it possible to use train_bit.py for processing multiple images? Additionally, I would like to ask if train_bit.py can deal with both sequential blurry images and out-of-order blurry images?

[lxin98]

The training code is also only for reasoning about three images, right?

[zzh-tech]

In the training and inference stage, the model only supports 3 nearby images as inputs.
As for out-of-order and more than 3 blurred images, the model needs to be retrained or its structure needs to be modified.