search

Algolzw / BSRT

Pytorch code for "BSRT: Improving Burst Super-Resolution with Swin Transformer and Flow-Guided Deformable Alignment", CVPRW, 1st place in NTIRE 2022 BurstSR Challenge (real-world track).
MIT License
179 stars 13 forks source link

只在Real data上训练 #1

Open greatlog opened 2 years ago

greatlog commented 2 years ago

作者您好，请问您有没有试过只在Real data上训练呢？结果怎么样？是否会出现过拟合的现象？

Algolzw commented 2 years ago

会过拟合，且效果是肯定没有finetune好的，具体的可以参考我们去年的BurstSR论文 EBSR, Table 1.