- Bring Ramp and Yolo v8 in this dockerized individual implementation from https://github.com/hotosm/fAIr-utilities/ , we would need to develop metadata , model card for both of the model , use same kind of accuracy representation if possible ( IOU ) , Both of the model should be indepedently implemented in a way sep docker container can be executed from the metadata
We would need to be able to run
- Run preprocessing
- Training
- Inference
- Postprocessing
from the metadata specified and always pass the same input and get the output as point / line / polygon , rest of the things can be burried inside as long as they are STAC compliant !
We would need to be able to run
from the metadata specified and always pass the same input and get the output as point / line / polygon , rest of the things can be burried inside as long as they are STAC compliant !