This week I completed training a new model using the new dataset which gave me an accuracy of up to 93.2%. We also decided to remove the Jetson and use the Raspberry Pi. So I have been working on integrating the Raspbveery Pi into our IOS app as well as host the module. I have also been working on the final presentation due tomorrow.
My progress is slightly behind schedule as a result of the prior Jetson issues, but I hope to get back on schedule soon.
Next week, I hope to finish integrating the components and conduct End-to-End testing.
As I have worked through this project, one of the biggest things I have learned is to find different ways to analyze an issue. More specifically, when the model wasn’t training properly, I had to experiment with various testing parameters such as alpha, batch size, and epochs as well as adding an additional layer of testing while training the model to see if the model was overfitting or underfitting with the data.