Pytorch Implementation of two ICLR papers
- Training Recurrent Neural Networks Online by Learning Explicit State Variables. Somjit Nath, Vincent Liu, Alan Chan, Xin Li, Adam White and Martha White.
- Unbiased Online Recurrent Optimization. Corentin Tallec, Yann Ollivier.
python main_torch.py --dataset cw --data_size 10000 --model fpp --state_update True --overlap True --T 10 --num_update 1 --batch_size 1 --lr 0.001 --buffer_size 100 --num_run 10
You can see the optional arguments by running:
python main_torch.py --help