Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore how to scale large language models and multimodal architectures efficiently using NVIDIA NeMo Framework Megatron-LM on Databricks in this 21-minute conference talk. Discover how this open source solution integrates GPU acceleration and advanced parallelism with Databricks Delta Lakehouse to streamline workflows for pre-training and fine-tuning models at scale. Learn about context parallelism, a unique NeMo capability for parallelizing over sequence lengths that makes it particularly effective for video datasets with large embeddings. Examine the real-world case study of TwelveLabs' Pegasus-1 model to understand how NeMo enables scalable multimodal AI development spanning from text to video processing. Gain insights into establishing new standards for LLM workflows through efficient data management and computational power optimization. The session is presented by Aiden Lee, Chief Technology Officer & Co-Founder of Twelve Labs, Inc, and Wenwen Gao, Product Manager at NVIDIA, providing both industry and technical perspectives on advanced model development practices.
