Getting Started with Visual Studio Tools for AI
- 3. Integrated w/Azure Machine Learning Integrated w/Cognitive Services Create new deep learning projects easily Scale Out with Azure Batch AI Generate C# code from TensorFlow & ONNX models Convert models to ONNX Monitor model training progress & GPU utilization Visualize your model training with TensorBoard Get started quickly with the Samples Gallery Visual Studio Tools for AI AI developer tool to train models & infuse AI into your apps http://aka.ms/vstoolsforai http://aka.ms/vscodetoolsforai
- 17. Creating Art with Image Style Transfer Models Data Sources Ingest / Prepare Model Train with Cloud AI Deploy Consume AC TION INTELLIGENC EDATA Azure Blob Raw storage Azure Machine Learning Docker Image + IoT Hub Model Update + Manageability 10 01 Model: VGG-19 Code: Tensorflow and Keras Microsoft Common Objects in Context (COCO) 328k images 91 different types of objects could be recognized by a 4 year old Visual Studio Tools for AI Deep Learning Virtual Machine (DLVM)
- 18. • We want to preserve stylistic features but not spatial structure • Loss functions measure high-level perceptual and semantic differences between images • Use pretrained loss network called VGG-19 trained on the ImageNet dataset Creating Art with Image Style Transfer Models Inspired by “Perceptual Losses for Real-Time Style Transfer and Super-Resolution” https://arxiv.org/abs/1603.08155 Approach (see paper for more detail) What’s happening inside
- 19. Create new deep learning projects easily Use TensorFlow, CNTK, Keras, Caffe2, Chainer, and more. Debug and iterate quickly with the power of Visual Studio.
- 20. Integrated with Azure Machine Learning Get started quickly with the sample gallery. Manage your experiments and models. Deploy in the cloud or on the edge.
- 21. Scale Out with Azure Batch AI Elastically scale training in Azure. Select Docker container, # VMs per job. Pay only for what you use when jobs are running.
- 22. Monitor model training progress & GPU utilization Integrated with TensorBoard to easily monitor & visualize your model training. GPU heatmap provides visibility for optimizing your resource utilization
- 23. Storage Browser to upload data, copy model & view logs Easily upload data to remote machines, download model files and view logs Working in the cloud is as convenient as your desktop.
- 24. Infuse AI into your apps today Include model in your app like any other resource … Or deploy to Azure ML and call via REST API
- 26. Bringing the best of AI to Azure and the best of Azure to AI Pre-Built AI Azure Cognitive Services Conversational AI Azure Bot Service Custom AI Azure Machine Learning
- 27. Integrated w/Azure Machine Learning Integrated w/Cognitive Services Create new deep learning projects easily Scale Out with Azure Batch AI Generate C# code from TensorFlow & ONNX models Convert models to ONNX Monitor model training progress & GPU utilization Visualize your model training with TensorBoard Get started quickly with the Samples Gallery Visual Studio Tools for AI AI developer tool to train models & infuse AI into your apps http://aka.ms/vstoolsforai http://aka.ms/vscodetoolsforai
- 28. find out more! Learn more about Cognitive Services – https://aka.ms/cogsvcs Azure ML – https://aka.ms/azuremlbuild Visual Studio Tools for AI - https://aka.ms/vstoolsforai Deep Learning VMs - https://aka.ms/dlvmbuild Batch AI – https://aka.ms/batchaibuild ML.Net - https://docs.microsoft.com/en-us/dotnet/machine-learning/