We have decided to stop using Git LFS and migrate our model weights to AWS S3.
When km_predict finds that the weights for the configured model architecture are missing, it would first download the weights and then start with the sub-tiling and segmentation as usual.
We have decided to stop using Git LFS and migrate our model weights to AWS S3.
When km_predict finds that the weights for the configured model architecture are missing, it would first download the weights and then start with the sub-tiling and segmentation as usual.