Building a Powerful Pet Detector in Notebooks

Learn how to build a pet breed detector using deep learning in this EuroPython 2019 conference talk. Explore the process of training, optimizing, and deploying a convolutional neural network model using Azure Notebooks and TensorFlow. Discover transfer learning techniques to retrain a MobileNet model on the Oxford IIIT Pet Dataset for recognizing dog and cat breeds. Gain insights into hyperparameter tuning to improve model accuracy and learn how to deploy the finished model as a web service. Follow along as the speaker demonstrates creating an accurate image recognition model using simple yet powerful techniques, covering topics such as deep learning concepts, machine learning workflows, and Azure ML Service.