The CNN and MLP models initially included hyperparameter tuning as well, however due to Google Colab having limited time on GPU usage and just taking too long or breaking by using their default CPU resources, I decided to remove it from the code, but still wanted to note it down and present some of the unfinished results it provided.
