Trainable parameters vs information in training images

I want to train the ResNets 50 model in the Assignment to process a “local” set of images, for a different categorical classification (10 classes, but I don’t think the details should matter at this point).
The total training set that I have consists of about 18 million pixels, grayscale, byte resolution (0-255). Given that the ResNets 50 has 23,546,886 trainable parameters, which is more than the total pixels in the training set, won’t that already announce overfitting? (Overparametrization as we would say in modeling).
Thanks for your help

Hi, Eduardo.

I don’t think there is any mathematical relationship that would say anything about overfitting being implied by the number of trainable parameters being greater than the number of features in the input. How complex are the features that you are trying to discern in the input images?