2

I am implementing a RL agent with policy gradient method. I define a dense network for actor and another dense network for critic. For example, my critic network is:

state_input = Input(shape=(self.num_states,))
x = Dense(self.hidden_size, activation='tanh')(state_input)
for _ in range(self.num_layers - 1):
    x = Dense(self.hidden_size, activation='tanh')(x)

out_value = Dense(1)(x)

model = Model(inputs=[state_input], outputs=[out_value])
model.compile(optimizer=SGD(lr=self.learning_rate), loss='mse')

In training phase I'm calling tensorboard:

from keras.callbacks import TensorBoard

tensorboard = TensorBoard(log_dir="/logs/{}".format(time()), 
                          histogram_freq=1, batch_size=32, 
                          write_graph=True, write_grads=True,
                          write_images=True, embeddings_freq=0, 
                          embeddings_layer_names=None, 
                          embeddings_metadata=None, 
                          embeddings_data=None, update_freq='epoch')

critic_loss = self.critic.fit([obs, advantage, old_prediction], [action], 
                                 batch_size=self.batch_size,
                                 shuffle=True, epochs=self.epochs, verbose=False,
                                 callbacks=[tensorboard_actor])

But I'm getting this error:

TypeError: 'module' object is not callable
4

1 回答 1

4

TypeError：在您的情况下，“模块”对象不可调用是由时间模块引起的

我假设您将时间模块导入为

import time

并调用函数 time()

tensorboard = TensorBoard(log_dir="/logs/{}".format(time())

可以通过导入轻松解决：

from time import time
于 2019-10-07T22:29:39.790 回答