Table of Contents¶
- Get the data for analysis
- Train an image classification model
- Deploy model
Deep Learning models are huge and requires high computation for inferencing. Can we train Deep Learning models which require less computation power, are smaller in size and can be deployed on mobile phones? Well, the answer is 'yes'. With the integration of capability to train TensorFlow lite models with ArcGIS API for Python, we can now train DL models that can be deployed on mobile devices and are smaller in size.
Where can we use them? We can use them up to train multiple DL models to perform classification tasks specifically for mobile devices. One such integration we did is in the "Survey123" application which is a simple and intuitive form-centric data gathering solution being used by multiple surveyors while performing ground surveys, where we integrated a tf-lite model to classify different plant species while clicking it's picture in the app.
This notebook intends to showcase this capability to train a deep learning model that can be used in mobile applications for a real time inferencing using TensorFlow Lite framework. As an example, we will train the same plant species classification model which was discussed earlier but with a smaller dataset.