Case Study
How Krea AI Accelerates Foundation Model Training with WEKA
Krea AI trains state-of-the-art foundation models for 30M+ users using WEKA NeuralMesh™, achieving ~100% GPU utilization and ~40% MFU at petabyte scale.
Krea AI trains state-of-the-art foundation models for 30M+ users using WEKA NeuralMesh™, achieving ~100% GPU utilization and ~40% MFU at petabyte scale.
