The purpose of this library is to build a mathematics classifier. You'll find scripts that:
- scrape and preprocess data from the arXiv
- build and train RNN models using TensorFlow
- build and train Unigram models
- compute statistical anaysis of text (e.g., TFIDF)
using our Cornetto library (see modules
).
First you need to create the training data. Use the class RNNTrainingData
from the data_handlers.py
module.
rnn_training_data = RNNTrainingData.build([...])
See data_handlers
for more details on the build
method. This gets passed directly to our RNNModel
class from rnn_model.py
.
model = RNNModel(model_params)
model.fit(rnn_training_data)
You can save and load models with model.save([...])
and model.load([...])
.