- A MultiStar model can be trained by running main.py, where parameters, datasets, etc. are also specified.
- The necessary dependencies can be installed as anaconda environment with conda_environment.yml.
- For the UNet backbone, the confnets implementation (https://github.com/imagirom/ConfNets) is used.
- The datasets are generated with the files in datagen/.
rikuturkki / multistar Goto Github PK
View Code? Open in Web Editor NEWThis project forked from overlapping-instances/multistar