llmpass / RSTT

Official pytorch implementation of paper "RSTT: Real-time Spatial Temporal Transformer for Space-Time Video Super-Resolution"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Train on myself datasets

jackxx33 opened this issue · comments

Hello,I am very interested in your work. I want to use this model to train my own dataset,is this possible?

Definitely possible, just prepare the lmdb in the same way: 4 frame input 7 frame output.

Does this model have to 4 frame input 7 frame output? Can I change some places to achieve 2x or 3x frame interpolation? And why the num_frames in 'RSTT-S.yml' is 7 instead of 4? Thanks!

Yes. It can achieve 2x or 3x frame interpolation, but you have to change the code to feed the right number of output frames. Then, change the code starts at

if i % 2 == 0:
.