Detect Sarcasm in news headlines. Refer to the .ipynb file for the entire code. Dataset linked here.
Model trained using TensorFlow v2.2.0
model.summary()
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding (Embedding) (None, 32, 32) 320000
_________________________________________________________________
dropout (Dropout) (None, 32, 32) 0
_________________________________________________________________
conv1d (Conv1D) (None, 28, 64) 10304
_________________________________________________________________
max_pooling1d (MaxPooling1D) (None, 7, 64) 0
_________________________________________________________________
lstm (LSTM) (None, 7, 128) 98816
_________________________________________________________________
dropout_1 (Dropout) (None, 7, 128) 0
_________________________________________________________________
lstm_1 (LSTM) (None, 64) 49408
_________________________________________________________________
dropout_2 (Dropout) (None, 64) 0
_________________________________________________________________
dense (Dense) (None, 32) 2080
_________________________________________________________________
dropout_3 (Dropout) (None, 32) 0
_________________________________________________________________
dense_1 (Dense) (None, 16) 528
_________________________________________________________________
dropout_4 (Dropout) (None, 16) 0
_________________________________________________________________
dense_2 (Dense) (None, 8) 136
_________________________________________________________________
dropout_5 (Dropout) (None, 8) 0
_________________________________________________________________
dense_3 (Dense) (None, 1) 9
=================================================================
Total params: 481,281
Trainable params: 481,281
Non-trainable params: 0
_________________________________________________________________