I solved the problem with the image size last night and I should thank you for that, however, after I trained the network for 24 epochs, the prediction of custom images are still far from correct. The percentage of each label it gives for custom images is around 25%, while testing random images from the data set gives very definite answer (50% for one label). Is it because of overfitting or not enough epochs? Is there any advice to improve the correctness for custom images?
I solved the problem with the image size last night and I should thank you for that, however, after I trained the network for 24 epochs, the prediction of custom images are still far from correct. The percentage of each label it gives for custom images is around 25%, while testing random images from the data set gives very definite answer (50% for one label). Is it because of overfitting or not enough epochs? Is there any advice to improve the correctness for custom images?