To run the shared head experiment on the Lifelong Text Classification benchmark:
python main.py --max_train_size=115000 --head_mode=uni --task_stream="domainshift"
To run the multiple head experiment on the Lifelong Text Classification benchmark:
python main.py --max_train_size=115000 --head_mode=multi --task_stream="domainshift"