这是我的功能:
scores = dqn.test(env, nb_episodes=1, visualize=False, verbose=1, callbacks=[CustomCallback()])
我从https://www.tensorflow.org/guide/keras/custom_callback
我有一个用于训练我的 DQN 的标准 openai 健身房环境:
class TestEnv(Env):
def __init__(self):
...
def step(self, action):
...
return sum(self.state.values()), reward, done, info
def render(self):
pass
def reset(self):
...
return sum(self.state.values())
如果可能的话,我想从回调类中绘制sum(self.state.values())的值。我想了解如何将这些值传递给方法中的回调类:
class CustomCallback(keras.callbacks.Callback):
...
def on_epoch_end(self, epoch, logs=None):
keys = list(logs.keys())
print("End epoch {} of training; got log keys: {}".format(epoch, keys))
print( **sum(self.state.values())**)