Formas incompatibles: [128,37] frente a [128,34]

He añadido una capa de atención en un modelo LStM para codificador-decodificador.

La función model.fit

history = model.fit_generator(generator = generate_batch(X_train, y_train, batch_size = batch_size),
                    steps_per_epoch = train_samples//batch_size,
                    epochs=epochs,
                    validation_data = generate_batch(X_test, y_test, batch_size = batch_size),
                    validation_steps = val_samples//batch_size)

Y este es el error que obtengo

---------------------------------------------------------------------------
InvalidArgumentError                      Traceback (most recent call last)
<ipython-input-42-dc64566948be> in <module>()
      3                     epochs=epochs,
      4                     validation_data = generate_batch(X_test, y_test, batch_size = batch_size),
----> 5                     validation_steps = val_samples//batch_size)

9 frames
/usr/local/lib/python3.6/dist-packages/tensorflow/python/eager/execute.py in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
     58     ctx.ensure_initialized()
     59     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
---> 60                                         inputs, attrs, num_outputs)
     61   except core._NotOkStatusException as e:
     62     if name is not None:

InvalidArgumentError:  Incompatible shapes: [128,37] vs. [128,34]
     [[node metrics_3/acc/Equal (defined at /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3009) ]] [Op:__inference_keras_scratch_graph_19367]

Function call stack:
keras_scratch_graph

Mi tamaño de lote es 128.

La función generar lote es

def generate_batch(X = X_train, y = y_train, batch_size = 128):
    ''' Generate a batch of data '''
    while True:
        for j in range(0, len(X), batch_size):
            encoder_input_data = np.zeros((batch_size, max_length_src),dtype='float32')
            decoder_input_data = np.zeros((batch_size, 34),dtype='float32')
            decoder_target_data = np.zeros((batch_size, max_length_tar, num_decoder_tokens),dtype='float32')
            for i, (input_text, target_text) in enumerate(zip(X[j:j+batch_size], y[j:j+batch_size])):
                for t, word in enumerate(input_text.split()):
                    encoder_input_data[i, t] = input_token_index[word] # encoder input seq
                for t, word in enumerate(target_text.split()):
                    if t<len(target_text.split())-1:
                        decoder_input_data[i, t] = target_token_index[word] # decoder input seq
                    if t>0:
                        # decoder target sequence (one hot encoded)
                        # does not include the START_ token
                        # Offset by one timestep
                        decoder_target_data[i, t - 1, target_token_index[word]] = 1.
            yield([encoder_input_data, decoder_input_data], decoder_target_data)

Aquí max_length_src = 34, max_length_tar = 37. El error parece deberse a esto.

Por favor ayuda.

de Agregar capa de atención al modelo Seq2Seq, este es su modelo (vuelva a intentar pegarlo y copiarlo):

ENCODER

num_encoder_tokens = 30
num_decoder_tokens = 10
latent_dim = 100

encoder_inputs = Input(shape=(None,))
enc_emb =  Embedding(num_encoder_tokens, latent_dim, mask_zero = True)(encoder_inputs)
encoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)
encoder_outputs, state_h, state_c = encoder_lstm(enc_emb)
# We discard `encoder_outputs` and only keep the states.
encoder_states = [state_h, state_c]

# Set up the decoder, using `encoder_states` as initial state.
decoder_inputs = Input(shape=(None,))
dec_emb_layer = Embedding(num_decoder_tokens, latent_dim, mask_zero = True)
dec_emb = dec_emb_layer(decoder_inputs)
# We set up our decoder to return full output sequences,
# and to return internal states as well. We don't use the
# return states in the training model, but we will use them in inference.
decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)
decoder_outputs, _, _ = decoder_lstm(dec_emb,
                                     initial_state=encoder_states)
decoder_dense = Dense(num_decoder_tokens, activation='softmax')
decoder_outputs = decoder_dense(decoder_outputs)

# Define the model that will turn
# `encoder_input_data` & `decoder_input_data` into `decoder_target_data`
model = Model([encoder_inputs, decoder_inputs], decoder_outputs)
model.summary()

DESCIFRADOR

# Set up the decoder, using `encoder_states` as initial state.
decoder_inputs = Input(shape=(None,))
dec_emb_layer = Embedding(num_decoder_tokens, latent_dim, mask_zero = True)
dec_emb = dec_emb_layer(decoder_inputs)
# We set up our decoder to return full output sequences,
# and to return internal states as well. We don't use the
# return states in the training model, but we will use them in inference.
decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)
decoder_outputs, state_h, state_c = decoder_lstm(dec_emb, initial_state=encoder_states)
attention = dot([decoder_outputs, encoder_outputs], axes=[2, 2])
attention = Activation('softmax')(attention)
context = dot([attention, encoder_outputs], axes=[2,1])
decoder_outputs = concatenate([context, decoder_outputs])
decoder_dense = Dense(num_decoder_tokens, activation='softmax')(decoder_outputs)

# Define the model that will turn
# `encoder_input_data` & `decoder_input_data` into `decoder_target_data`
model = Model([encoder_inputs, decoder_inputs], decoder_dense)
model.summary()

No tengo ningún problema para ajustar el modelo ...

model.compile('adam', 'categorical_crossentropy')

n_samples = 5
X_enc = np.random.randint(0,num_encoder_tokens, (n_samples,37))
X_dec = np.random.randint(0,num_decoder_tokens, (n_samples,34))
y = np.ones((n_samples,34,num_decoder_tokens))

model.fit([X_enc, X_dec], y, epochs=10)