metadata
license: apache-2.0
tags:
- medical-imaging
- segmentation
- medotter
MedOtter E2E zero predictor
This tiny public repository is the immutable acceptance fixture for MedOtter's
key-authenticated internal submission pipeline. Its top-level predictor.py returns
an all-background predictor matching the MedOtter SDK contract and therefore works
across the supported
semantic and instance segmentation adapters without model weights or downloads.
It is not a trained model, has no clinical utility, and exists only to verify model intake, pinned artifact resolution, worker execution, scoring, publication, and web display.