Adam-v2 medical model weights hit GitHub

Adam-v2 shifts medical AI from generic feature extraction to specialized hierarchical understanding, moving beyond simple image recognition toward structural anatomical awareness. Its three-branch architecture mirrors biological reality by understanding how anatomical parts form a whole, while pretrained ConvNeXt-B weights enable high-performance feature extraction for anomaly detection. Superior zero-shot transfer capabilities make it a versatile foundation for niche medical imaging, and the public availability of weights in the Eden repository accelerates development of anatomy-aware systems.