Microsoft expands its portfolio of machine learning tools
Microsoft has launched a new range of machine learning tools at its annual Ignite conference. With the new launch, the Windows maker wants to enable developers to utilise easy artificial intelligence (AI) based tools to build new advancements.
The new tools are the Azure Machine Learning Workbench, Azure Machine Learning Experimentation Service and the Azure Machine Learning Model Management Service. These are designed to help developers working on new AI models, as well as those who simply want to use pre-existing models.
The Azure Machine Learning Experimentation Service is designed to help developers deploy ML experiments. Microsoft has added support for open source frameworks like Caffe2, PyTorch, TensorFlow, CNTK and Cahiner. The experimentation service is designed to scale from local machines to hundreds of
GPUs in the cloud.
Machine Learning Workbench, on the other hand, is a desktop client for Windows and Mac. The tool can act as a control panel for your development life cycle. Lastly, the Azure Machine Learning Model Management Service uses Docker containers. Developers can manage and deploy their models to any Docker container using the service. The company has also included its own Kubernetes-based Azure Container Service within the learning workbench to enhance the coverage of its native development.
All the new machine learning developments are a part of Microsoft’s AI platform, which the company claims serves over 650,000 sessions per week.