Testing various architectures for the kaggle Toxic Comment Classification Challenge.
├── __init__.py
├── archs
│ ├── gru_max_avg_pool.py
│ ├── gru_max_avg_pool_attention.py
│ └── lstm_max_avg_pool.py
├── download_dataset.ipynb
├── gru_max_avg_pool_glove.840B.300d.ipynb
├── lstm_max_avg_pool_glove.840B.300d.ipynb
├── train.py
├── utils.py
└── weights
├── bigru.best.zip
├── bigru_simple.best.zip
├── bilstm.best.zip
├── bilstm_simple.best.zip
├── gru.best.zip
├── gru_simple.best.zip
├── lstm.best.zip
├── lstm_simple.best.zip
├── rnn_simple.best.zip
└── weights_base.best.zip