Training on custom data

JingyunLiang / MANet

Official PyTorch code for Mutual Affine Network for Spatially Variant Kernel Estimation in Blind Image Super-Resolution (MANet, ICCV2021)

https://arxiv.org/abs/2108.05302

Apache License 2.0

169 stars 23 forks source link

Training on custom data #22

Open rose-jinyang opened 1 year ago

rose-jinyang commented 1 year ago

Hello How are you? Thanks for contributing to this project. I know that this method is for train a blur kernel estimation model under on unsupervised mode. I am going to use this method for training a blur kernel estimation model on my custom data. My custom data contains ONLY low-quality images. I checked the option file "train_stage1.yml" for stage1 of MANet training but it contains the path to ONLY high resolution data folder so I'm confused.