Видео недоступно в Pусский. Видео доступно в English (US).
AI with Azure Machine Learning services: Simplifying the data science process
The process of developing machine learning models for production involves many steps. The data scientist must decide on a model architecture and featurization, they must train and tune these models, manage the compute resources required to execute and scale out training, manage the training data and make it available to the training compute, keep track of the different (hyper-) parameter combinations and model versions used along with the results they yielded. All that is often embedded in a complex flow to acquire and prepare the data on the one side and to post-process and deploy the model on the other.