hello, how to add changemixin when use bitemporal supervised

[csliuchang]

hello I have question about your repo:

1. how to add changeminxin when use bitemporal supervised, i see it in your paper table 4 but i cant find in codes?
2. could changestar use LEVIR-CD train Single-Temporal(another dataset is too big for train, i cant download it)
3. are your bitemporal suprvised methods just use torch.cat in the final layer?
   sorry for ask these question,

[Z-Zheng]

1. see here https://github.com/Z-Zheng/ChangeStar/blob/master/module/changestar_bisup.py
2. No, because there is no building label provided in LEVIR-CD. Maybe you can choose a small building segmentation dataset to try. But small dataset is difficult to guarantee the final performance under single-temporal supervision.
3. Bitemporal model simplifies the temporal swap module and uses torch.cat + many convs, not only the torch.cat.

[csliuchang]

THANKS BRO

[csliuchang]

i use the bitemporal suprvised methods for Industrial application，but i found the operation torch.cat may lost some small changes(this change is so important) of information that don't have obvious semantic features. so ask third question, dont mind, thanks again.