CNN Wafer Inspection Results

Model's accuracy vs Epochs' graphs showing the results obtained using a learning rate of 0.0001, the adam optimizer, and 128 neurons in a fully connected layer, for:

1. Categorical Classification for the entire dataset
2. Categorical Classification for the entire dataset (to distinguish between the different failure modes and not-failed), and additionally using data augmentation - horizontal and vertical flips
3. Categorical Classification just for the failed wafers (to distinguish between the different failure modes only), and using data augmentation
4. Binary Classification for the entire dataset to tell whether a wafer has failed or not, also using data augmentation
5. Binary Classification for equal parts of failed and non-failed wafers to tell whether a wafer has failed or not, once again using data augmentation