In [ ]:

    

import CNNModelCreatorWordLSTM as md

dimensions = 50
embedfile = 'glove.6B.' + str(dimensions) + 'd.txt'
authorList = [11, 18, 80, 88, 64]
doc_id = 1
chunk_size = 1000
nb_epoch = 30
glove = '../../glove/'
samples = 3200
batch_size = 10


    

In [ ]:

    

embeddings_index = md.readVectorData(embedfile, GLOVE_DIR=glove)


    

In [ ]:

    

(texts, labels, labels_index, samples) = md.loadAuthData(authorList, doc_id, chunk_size = chunk_size, samples = samples)


    

In [ ]:

    

(trainX, trainY, valX, valY) = md.preProcessTrainVal(texts, labels, chunk_size = chunk_size)


    

In [ ]:

    

embedding_matrix = None

embedding_matrix = md.prepareEmbeddingMatrix(embeddings_index, EMBEDDING_DIM = dimensions)


    

In [ ]:

    

model = md.compileModel(len(labels_index), embedding_matrix, chunk_size = 1000,
                       DROP_OUT = 0.5, EMBEDDING_DIM = dimensions)


    

In [ ]:

    

(model, history, train_acc, val_acc) = md.fitModel(model, trainX, trainY, valX, valY,
                                                   nb_epoch = 100, batch_size = batch_size)


    

In [ ]: