character-predictor

A python api for predicting the next word based on present text given. The character-predictor uses recurrent neural network(RNN) as prediction model for predicting the next character based on previous text.

Requirements

Its a python implementaion compatible with both 2 and 3.

Requirements:

sklearn <= 0.0
numpy <= 1.15.0
Theano <= 1.0.3

For adding to your project: Download the implementation to your project folder Add the lines to your file

> from char_pridect import PredictWord
> pword = PredictWord(model_file="model_file.save")

Use pword for predicting word e.g.

> text = "the partner of experi"
> pword = pword.predict(inp)
> pword
['experience ', 'experime ', 'experiencly ', 'experiend ', 'experiming ']

Training

For training character-prediction RNN model to specific text corpus or data.

if __name__ == '__main__':
    fname = 'pg.txt'
    cmodel = CharPredictNNModel(seq_len=32, hidden_lay_sz=(128,), model_file='model_file.save')
    cmodel.compile()
    cmodel.train(fname)

    # testing
    correct, total = cmodel.test(fname)

    sys.stdout.write("accuracy :", float(correct)/total)