Three scripts that can be uploaded directly as Kaggle kernels to make submissions.
quoradl_cv_averaging.py: performs cross validation and predicts on test set at each fold. Predictions are averaged over all folds.
quora_dl_with_features.py: two-legged neural network, where one leg processes the text with an RNN and the other processes the added features.
quoradl_embedding_ensemble.py: work in progress, a kernel that fits RNNs using multiple embeddings and trains an ensemble method on the results. We can add simple predictors based on features to this later.
Three scripts that can be uploaded directly as Kaggle kernels to make submissions.
quoradl_cv_averaging.py: performs cross validation and predicts on test set at each fold. Predictions are averaged over all folds.quora_dl_with_features.py: two-legged neural network, where one leg processes the text with an RNN and the other processes the added features.quoradl_embedding_ensemble.py: work in progress, a kernel that fits RNNs using multiple embeddings and trains an ensemble method on the results. We can add simple predictors based on features to this later.