Quickly deploy NVIDIA optimized generative AI on Azure AI Studio
summary
Shows how to take a model from Azure ML, run NVIDIA NIM to auto-optimize it, and deploy a working endpoint on Azure (NC A100 v4 with four A100 GPUs) that speaks the OpenAI API spec and supports streaming. Demo deploys Mixtral 8x7B via a simple YAML + two commands, returns successful completions and streaming tokens. Also deploys VISTA-3D, a MONAI medical CT model with 126 classes that works with few‑shot and interactive point‑click inference.
AI-written summary · about
chapters (14)
- 0:00NVIDIA AI on Azure AI Studio
- 1:15Enterprise AI deployment challenges
- 3:43NVIDIA NIM overview
- 6:31Deploying with NVIDIA NIM
- 9:43Demo setup and workspace
- 10:28Azure ML endpoint for Mixtral
- 11:30Deployment YAML for inference
- 12:57Testing model with Postman
- 14:44Chatbot integration
- 15:50VISTA-3D deployment demo
- 16:36VISTA-3D on Azure ML
- 18:04MONAI and VISTA-3D overview
- 19:09VISTA-3D segmentation demo
- 20:38Summary and resources
related talks
24mHow to build next-gen AI services with NVIDIA AI on Azure Cloud
28mHarnessing generative AI with NVIDIA AI and Microsoft Azure
30mNVIDIA AI Enterprise Registry for AzureML deployments of AI Workflows
59mLearn Live: Introduction to Azure OpenAI Service
48mUsing Azure AI Foundry to manage all your Large Language Models
59mBuild Next-Gen AI Apps with .NET and Azure
48mDeploy large language models responsibly with Azure AI
20mUnlocking Azure OpenAI: A 20-Minute Introduction