Empowering Generative AI with Databricks and AMD

Explore the critical role of GPU hardware in pretraining and fine-tuning generative AI models in this 42-minute conference talk. Discover how AMD's advanced GPU hardware and ROCm software are revolutionizing large-scale model training, moving away from a single-vendor-dominated industry. Learn to leverage AMD GPUs for processing private datasets, utilize Delta Sharing and Unity Catalog for seamless environment transitions, and implement serverless infrastructures for inference to aid monetization. Gain valuable insights into how Databricks and AMD are making Generative AI more accessible and cost-effective while maintaining robust governance and data privacy. Delve into comprehensive best practices for the complete training and deployment lifecycle of generative AI models, presented by Avinash Sooriyarachchi from Databricks and Keith Anderson from AMD.

Empowering Generative AI with Databricks and AMD