search

cs-giung / FSRNet-pytorch

PyTorch implementation of "FSRNet: End-to-End Learning Face Super-Resolution with Facial Priors" (https://arxiv.org/abs/1711.10703)
53 stars 8 forks source link

Evaluate the results of super-resolution on my images, #16

Open TeiZouen opened 1 year ago

TeiZouen commented 1 year ago

@cs-giung Thank you for your code. I would like to evaluate the results of super-resolution on my images, how do I evaluate them? Looking forward to your response.

cs-giung commented 1 year ago

Hello there! It's great to see your interest in this repository.

Regarding your inquiry: I believe it's possible to replace dataset with a custom dataset in eval.py (although I'm not entirely certain since it has been a while since I last worked on this).

However, it's important to note that this repository is no longer actively maintained, and the provided checkpoints may not meet the desired performance level. If your objective is to exclusively evaluate new images, I recommend exploring an alternative pre-trained model, such as the one available in the official repository at https://github.com/tyshiwo/FSRNet or considering other state-of-the-art models.

TeiZouen commented 1 year ago

@cs-giung
Thank you very much for your time and prompt reply. I will check eval.py and see how to replace my image. Next, I will try to reproduce the official repository you recommended. Thank you very much.