Thursday Response 7/15

1. Using the relevant parts from the entirety of the example script as a guide, fit a CNN model to your training data and validate using the beans dataset from Tensorflow datasets and then again train and validate using the eurosat dataset. Present your results and discuss the accuracy of each of model.

   • After running both datasets through the model we came across the following results and accuracies. The results were of some surprise to us. We were expecting the bean model to do better than the EuroSAT dataset beacuse we were expecting the beans to be a less complex dataset. This is because it has 3 classes compared to 10 with the EuroSAT. We tried fitting with both the val_ds and test_ds as the validaiton set and posted the results. The test did better because it was building off the results of the val_ds.

     Beans or EuroSAT	Validation	Test
     Beans	0.6796	.8041
     EuroSAT	0.7584	.8249

2. “Apply this same method to both the beans and eurosat datasets. Did your model performance improve? How many epochs were you able to run and how much time did it take? Present your results and discuss the accuracy of your augmented output for tomorrow’s class.”

   • Adding the augmentation resulted in the following results and accuracies. It was a drastic decrease in accuracy and made it so that the model took longer to train. We did 3 epochs for each because we saw that the model was not accurate or gaining overall accuracy. Each training took around 5 minutes for the bean dataset and 3 minutes for the EuroSAT dataset. The results with the augmentation followed our inclination more that the beans model would be better than the EuroSAT model.

     Beans or EuroSAT	Validation	Test
     Beans	.2987	.3507
     EuroSAT	.2436	.3123