This repository is the official implementation of Paper Title.
Install package and dependencies:
python setup.py install
pip install -r requirements.txt
To generate Penn Tree Bank dataset, Anna dataset or a Toy dataset run:
python cprnn/features/build_features.py -d [toy/ptb/anna]
Enter training configs in configs.yaml
then run the command below. Results will be saved in runs
folder.
python train.py
To visualize the training process run:
tensorboard --logdir=runs
To run all experiments in the figure below run:
bash job.sh
Then visualize using:
python cprnn/visualization/bpc_vs_params.py
CP-RNN Rank 128 @ Epoch 50
The took said a come in investment it were all admines in the coming thrual and spock industrial stock o
CP-RNN Rank 16 @ Epoch 50
Then they in monced this and the as anessue otton thestences a n't as big sayernment agricens is costsha