The company has launched three major tools for developers: The Azure Machine Learning Experimentation Service, The Azure Machine Learning Workbench, and the Azure Machine Learning Model Management Service.
In addition, Microsoft has also launched new set of tools for developers and non-developers. For developers, they have made Visual Studio Code IDE for making models with CNTK, TensorFlow, Theano, Keras and Caffe2. Azure based machine learning models has brought up by Microsoft for non-developers especially for Excel users.
The most interesting announcement made by Microsoft’s Joseph Sirosh notes was the tool’s feature to transform your data so that machine learning algorithm can work automatically.
Microsoft enables developers to train quickly and deploy machine learning experiments by the experimentation service. All the open source frameworks are supported by this service and it can scale from large number of GPU’s in the clouds to a local machine. You may also like to hear that Apache Spark is also supported by this tool on Azure HDInsight Clusters. Developers get full versioning of their experiments by keeping track of all the models, configurations and data through this service.
The core point of the announcements made by Microsoft is that Developers will keep receiving the expansion of toolbox that will help them making machine-learning based applications- both for inner and outer customers. And among all the interesting announcements, the most lucrative one is that these tools support many non-Microsoft frameworks. Microsoft has brought a platform to support all kind of frameworks. Few years back, every framework has its own advantages and disadvantages so, any tool cannot be applicable for all kind of framework. Thankfully, Microsoft took initiative and offered a magnificent tools to developers.