transformer_35.py

encoder = Encoder(en_vocab_size, MODEL_SIZE, NUM_LAYERS, H)
decoder = Decoder(fr_vocab_size, MODEL_SIZE, NUM_LAYERS, H)

NUM_EPOCHS = 100

start_time = time.time()
for e in range(NUM_EPOCHS):
    for batch, (source_seq, target_seq_in, target_seq_out) in enumerate(dataset.take(-1)):
        loss = train_step(source_seq, target_seq_in,
                          target_seq_out)

    print('Epoch {} Loss {:.4f}'.format(
          e + 1, loss.numpy()))

    if e % 10 == 0:
        end_time = time.time()
        print('Average elapsed time: {:.2f}s'.format((end_time - start_time) / (e + 1)))
        try:
            predict()
        except Exception as e:
            print(e)
            continue