wyolo is the definitive framework for scaling YOLOv8 and RT-DETR training. Seamlessly integrated with MLflow, DVC, and S3, it transforms computer vision experiments into industrial-grade pipelines.
Designed for organizations that need to manage hundreds of models with total traceability.
Automatic logging of metrics, hyperparameters, and artifacts. One-click model registration in the enterprise registry.
Track every byte of your training data. Full reproducibility guaranteed through native Data Version Control integration.
Intelligent GPU memory detection. Automatically scales batch sizes to maximize throughput without OOM errors.
Robust worker implementation that handles dataset checks, GPU verification, and training cycles with error recovery.
Centralized artifact management. Sync your weights, plots, and results to secure enterprise storage automatically.
Powerful command-line interface for CI/CD integration and hyperparameter tuning sweeps.
Initialize a trainer with your enterprise config and let wyolo handle the complexities of MLOps lifecycle.
wyolo is part of the Wisrovi AI suite, designed to bridge the gap between academic research and industrial computer vision implementation.
Visit wisrovi.dev