Project that aims to train a multitasking model where each task has their own dataset rather than a sample contains multi-label.
For convenient and bug-free, please use conda virtual enviroment then execute the following commands:
$ pip install -r requirements.txt