Validating while training?

hengchuan / RDN-TensorFlow

A TensorFlow implementation of CVPR 2018 paper "Residual Dense Network for Image Super-Resolution".

MIT License

143 stars 47 forks source link

Validating while training? #19

Open wen04141321 opened 5 years ago

wen04141321 commented 5 years ago

Hello, I want to ask if you don't need to train and validate when you do the experiment? So how do you find the best model?

hengchuan commented 5 years ago

we pay more attention to the visual effects rather than psnr in daily work, which is hard to be validated. surely validating is necessary in analysis, I will add it to the scripts.

Pichairen commented 5 years ago

Without validating,I have no idea how many epoch should be set and how to avoid overfiting while training? any help?

Pichairen commented 5 years ago

@wen04141321 I tried validating when training, depart some data from training set as validation set ,and make eval.h5 for your validation set, when training ,you can add code for validating with feed_dict your validation data.