I have to deliver a project to my university which is based on Fast AI's GAN lesson. I have a few notebooks in Google Colab and the code is finished 95%. I need someone to test different parameters and maybe find some better datasets or modify the current one (DIV2K) to get better results as I don't have much time given I have a full-time job.
GAN is supposed to enhance quality of images. Critic and Generator for the GAN are pretrained and the Generator (U-net) is also used separately (to compare results later) so I'd need to get both GAN and U-net to have decent results. All images are saved on my Google drive, except for the GAN notebook but I was going to change it.
So here you have the four notebooks I use:
1) [login to view URL] - Making low resolution images out of DIV2K dataset for training
2) [login to view URL] - U-net
3) [login to view URL] - GAN
4) [login to view URL] - preprocessing images before getting predictions, basically making then square as they will be later uploaded to a webpage which requires them to be square
8 freelanceria on tarjonnut keskimäärin 106$ tähän työhön
NN ( neural network ) & GA ( genetic algorithm like a human Brain that work as MAT LAb, I have done this type of work. if any inquiry whatsaap 9918348699