This repo is an exmaple of
Text CNN for sentiment analysis
with Keras and keras.dataset.imdb.
$ make train
Epoch 1/5
176/176 [==============================] - 26s 147ms/step - loss: 0.2090 - acc: 0.6815 - val_loss: 0.1584 - val_acc: 0.8036
Epoch 2/5
176/176 [==============================] - 23s 131ms/step - loss: 0.1422 - acc: 0.8238 - val_loss: 0.1455 - val_acc: 0.8028
Epoch 3/5
176/176 [==============================] - 23s 130ms/step - loss: 0.1147 - acc: 0.8603 - val_loss: 0.1272 - val_acc: 0.8332
Epoch 4/5
176/176 [==============================] - 23s 131ms/step - loss: 0.0846 - acc: 0.9054 - val_loss: 0.1295 - val_acc: 0.8224
Epoch 5/5
176/176 [==============================] - 23s 131ms/step - loss: 0.0539 - acc: 0.9476 - val_loss: 0.1317 - val_acc: 0.8288
$ make test
loss: 0.12788543040275574
acc: 0.83412