Hi Mikhail! Thank you for the feedback! For now, the assignments are designed not to accept modifications outside the ### START CODE HERE ### and ### END CODE HERE ### tags (you can find other troubleshooting tips here). We know this can be improved and we will definitely consider this in future revisions. In the meantime, kindly revert the training cell to this then revise your loss to work with it:
model = create_model(NUM_WORDS, EMBEDDING_DIM, MAXLEN)
history = model.fit(train_padded_seq, train_label_seq, epochs=30, validation_data=(val_padded_seq, val_label_seq))
Thank you for bringing this up!