Increase accuracy for Pytorch implementation

I need someone who helps me in increasing accuracy for one Pytorch based application. I will give you access to my private GitHub repo once I award it. You need to add some document or readme file about what you have changed. I will give more details in the message. You don’t have to write too many lines of code, just a few lines may be. You just have to increase the accuracy by fixing the CNN might be. It’s a small project and I need to finish ASAP. My max budget is $15. I cannot increase more than that. Let’s talk.