kamillasak

Results 4 comments of kamillasak

I manage to use GPU instad of CPU, i make 100 epochs. Results unfortunatelly doesnt look good. ![1](https://user-images.githubusercontent.com/65394917/172332823-8f0b4651-684f-4976-b546-3fb0998ed935.jpg) ![2](https://user-images.githubusercontent.com/65394917/172332826-0ccd7516-978c-4e0b-acb6-fa81d26e90bc.jpg) ![3](https://user-images.githubusercontent.com/65394917/172332830-3c6022ac-127d-466c-b20d-d81306d7587b.jpg) ![4](https://user-images.githubusercontent.com/65394917/172332832-469c7214-25a9-4d90-84ad-1024a2a1fb75.jpg) Any ideas whats wrong? Edit: Graphics card is RTX3060,...

User parameters for pokemon .jpg data are: ![image](https://user-images.githubusercontent.com/65394917/172536093-45134751-9802-428c-aeff-fdd7463f2a38.png) And my conda environment is conda create --name dlss tensorflow-gpu=1.14.0 keras=2.3.1 scikit-image scikit-learn numpy matplotlib pillow Edit: On google collab it seems...

I have tried the newest versions of all libraries, and same results. I will look on these links, but im looking for supersampling mainly. Edit: on google collab i get...