AI/ML Experiment Tracking and Visualization Platform
3.9k 2026-04-26
SwanHubX/SwanLab
SwanLab is an open-source, modern-design platform for tracking, visualizing, and analyzing AI/ML training experiments, supporting cloud and self-hosted deployments.
Core Features
Comprehensive AI/ML experiment tracking and data visualization.
Flexible deployment options: Cloud and Self-hosted (including Kubernetes).
Deep integration with 50+ mainstream ML frameworks (e.g., PyTorch, Keras, Transformers).
Advanced dashboard with features like LightningBoard, experiment pinning, and baseline comparison.
Team collaboration, project management, and hardware monitoring (GPU, AMD ROCm).
Detailed Introduction
SwanLab is a professional, modern-designed AI training analysis platform tailored for model training teams. It streamlines the entire machine learning experiment lifecycle by offering robust tracking, visualization, and analysis capabilities. With seamless integration across over 50 mainstream frameworks, SwanLab helps researchers and engineers easily log metrics, visualize results, compare experiments, and collaborate effectively, whether deployed in the cloud or on-premises, ultimately accelerating the discovery of optimal model parameters.