because of low quality of data set, it can't use to generate high resolution image.
just use it to fine-tune the action in lower weight
maybe under 0.8
in the next version，I will find or create higher data set to train
because of low quality of data set, it can't use to generate high resolution image.
just use it to fine-tune the action in lower weight
maybe under 0.8
in the next version，I will find or create higher data set to train