It will extract the sentences and classes from the dataset, split it into train/val/test and save it in a convenient format for model.
Build vocabularies and parameters for dataset by running
python build_vocab.py --data_dir data/
It will write vocabulary files chars.txt and classes.txt containing the amino acid notations and classes in the dataset. It will also save a dataset_params.json with some extra information.
Train Simply run
python train.py --data_dir data --model_dir experiments/base_model
It will instantiate a model and train it on the training set following the hyperparameters specified in params.json. It will also evaluate some metrics on the development set.
First hyperparameters search Created a new directory learning_rate in experiments. Now, run
python search_hyperparams.py --data_dir data --parent_dir experiments/learning_rate
It will train and evaluate a model with different values of learning rate defined in search_hyperparams.py and create a new directory for each experiment under experiments/learning_rate/.
Display the results of the hyperparameters search in a nice format
Evaluation on the test set Run many experiments and selected best model and hyperparameters based on the performance on the development set,finally evaluate the performance of model on the test set. Run
python evaluate.py --data_dir data --model_dir experiments/base_model
请发表评论