The Critical Role of Storage in Optimizing AI Training Workloads

Explore a comprehensive technical presentation that examines how storage solutions critically impact AI workload optimization, particularly focusing on storage-intensive AI training workloads. Learn about the intricate interactions between AI models and storage systems during training, with detailed explanations of data loading and checkpointing mechanisms. Discover how AI frameworks, specifically PyTorch, leverage various storage connectors to interface with different storage solutions. Gain practical insights into the implementation of file-based storage and object storage within AI training environments. Presented by industry experts Ugur Kaynar from Dell Technologies and Jayanthi Ramakalanjiyam from Celestica, master the fundamental concepts of storage architecture in AI systems, understand the technical requirements for efficient AI model training, and learn best practices for optimizing storage solutions in AI workflows.