fannymonori / tf-espcn Goto Github PK

If possible, can you please put up a colab implementation of ESPCN model measuring the inference time of a video or image?

Hi I am training a new model with my own data set. I was wondering if there was a "best practices" when it comes to data?

Should all my training data be the same resolution?
Should all my training data be the highest resolution possible? The lowest? a mix?
Can I continuously add data to and retrain my model?

Thank you!!

Thank you for sharing your code.

I trained the model with my own data.
training was good.
However, errors occur when testing the model.
The test works fine on Scale 2, but on 3x and 4x

HR_image_YCrCb = np.concatenate((Y, Cr, Cb), axis=2) in run.py line 146
The following error occurs.

ValueError: all the input array dimensions for the concatenation axis must match exactly, but along dimension 0, the array at index 0 has size 278 and the array at index 1 has size 417

How do I fix it?

please help me.

Hi, I wonder if someone might know why I get blurriness comparable to bicubic?

I've posted results here: https://forum.opencv.org/t/why-so-blurry-please/11069/2

Works well with the sample images such as this one, but not with images that are anti-aliased.

this one works well:
https://miro.medium.com/max/562/1*7sqZ6SIRlyR6ex1IbFjtFg.webp

Hi, I'm going here from your other code project!
the same problem is how to test the images with an img_file?
Thanks!