search

Doodleverse / segmentation_gym

A neural gym for training deep learning models to carry out geoscientific image segmentation. Works best with labels generated using https://github.com/Doodleverse/dash_doodler
MIT License
45 stars 11 forks source link

remove DO_CRF option for post-processing estimated labels in `seg_images_in_folder` and install yml file #108

Closed dbuscombe-usgs closed 1 year ago

dbuscombe-usgs commented 1 year ago

The pydensecrf package is problematic and this has been well documented on these pages.

It is not a particularly useful option for post-processing. Plus, it prevents deployment on automated servers like Hugging Face because it is a dependency of doodleverse-utils

It should therefore be removed from all Zoo, Gym, and Doodleverse-utils functions, as well as CoastSeg

For Zoo, it involves modifying seg_images_in_folder.py and install yml file

dbuscombe-usgs commented 1 year ago

Fixed in https://github.com/Doodleverse/segmentation_gym/commit/8adb48dc51d4a9e705505665b9fbc47e1cf39f8a

Now needs doodleverse-utils version v0.0.14 https://pypi.org/manage/project/doodleverse-utils/release/0.0.14/

see also https://github.com/Doodleverse/doodleverse_utils/issues/16