Haiyang Mei comments

Results 9 comments of


                                            Haiyang Mei

I have the same problem.

I solved this problem! The input size of image should be 480x480.

I also can not find UCF dataset.

Sorry, the pre-trained model cannot be shared as it is too big. You can train the network from scratch.

We will release the dataset and pre-trained model in the near future.

Yes.

I encounter the same problem!

Withou initialized model, the train loss is 3.86349e+07 and can't fall. Can you provide the initialized model? Thanks.

How could you increase image resolution ?Where is the sub-pixel which mentioned in paper ESPCN ?