用tensorflow实现强化学习的dql算法报错：tensorflow.python.framework.errors_impl.InternalError: Could not find valid device for node. Node: {{node OneHot}} = OneHot[T=DT_FLOAT, TI=DT_FLOAT, axis=-1](dummy_input, dummy_input, dummy_input, dummy_input)

代码：

import tensorflow as tf

import numpy as np

import gym

import random

from collections import deque

tf.enable_eager_execution()

num_episodes = 500

num_exploration_episodes = 100

max_len_episode = 1000

batch_size = 32

learning_rate = 1e-3

gamma =1

initial_epsilon = 1

final_epsilon = 0.01

class QNetwork(tf.keras.Model):

    def __init__(self):

        super().__init__()

        self.dense1 = tf.keras.layers.Dense(units= 24, activation = tf.nn.relu)

        self.dense2 = tf.keras.layers.Dense(units= 24, activation = tf.nn.relu)

        self.dense3 = tf.keras.layers.Dense(units = 2)

    def call(self, inputs):

        x = self.dense1(inputs)

        x = self.dense2(x)

        x= self.dense3(x)

        return x

    def predict(self, inputs):

        q_values = self(inputs)

        return tf.argmax(q_values, axis=-1)

env = gym.make('CartPole-v1')

model = QNetwork()

optimizer = tf.train.AdamOptimizer(learning_rate)

replay_buffer = deque(maxlen = 10000)

epsilon = initial_epsilon

for episode_id in range(num_episodes):

    state = env.reset()

    episode = max(

        initial_epsilon* (num_exploration_episodes - episode_id) / num_exploration_episodes,

        final_epsilon

    )

    for t in range(max_len_episode):

        env.render()

        if random.random() < epsilon:

            action = env.action_space.sample()

        else:

            action = model.predict(

                tf.constant(np.expand_dims(state,axis=0, dtype = tf.float32)).numpy()

            )

            action = action[0]

        next_state, reward, done, info = env.step(action)

        reward = -10. if done else reward

        replay_buffer.append((state,action,reward,next_state,done))

        state = next_state

        if done==1:

            print('episode %d, epsilon %f, score %d' % (episode_id, epsilon, t))

            break

        if len(replay_buffer) >=batch_size:

            batch_state, batch_action, batch_reward, batch_next_state, batch_done = [np.array(a, dtype=np.float32) for a in zip(*random.sample(replay_buffer, batch_size))]

            q_value = model(tf.constant(batch_next_state, dtype = tf.float32))

            y = batch_reward + (gamma* tf.reduce_max(q_value, axis=1))*(1-batch_done)

            with tf.GradientTape() as tape:

                loss = tf.losses.mean_squared_error(

                    lables=y,

                    predictions = tf.reduce_sum(model(tf.constant(batch_state)) * tf.one_hot(batch_action, depth=2), axis=1)

                )

            grads = tape.gradient(loss, model.variables)

            optimizer.apply_gradients(grads_and_vars = zip(grads,model.variables))

报错：/home/kalarea/.conda/envs/py35/bin/python /home/kalarea/PycharmProjects/dql_demo/dpq_cartplie.py

/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.

  from ._conv import register_converters as _register_converters

episode 0, epsilon 1.000000, score 12

2018-10-15 12:54:59.165871: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA

Traceback (most recent call last):

  File "/home/kalarea/PycharmProjects/dql_demo/dpq_cartplie.py", line 77, in

    predictions = tf.reduce_sum(model(tf.constant(batch_state)) * tf.one_hot(batch_action, depth=2),axis=1)

  File "/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/tensorflow/python/ops/array_ops.py", line 2439, in one_hot

    name)

  File "/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/tensorflow/python/ops/gen_array_ops.py", line 4563, in one_hot

    _six.raise_from(_core._status_to_exception(e.code, message), None)

  File "", line 3, in raise_from

tensorflow.python.framework.errors_impl.InternalError: Could not find valid device for node.

Node: {{node OneHot}} = OneHot[T=DT_FLOAT, TI=DT_FLOAT, axis=-1](dummy_input, dummy_input, dummy_input, dummy_input)

All kernels registered for op OneHot :

  device='CPU'; TI in [DT_INT64]; T in [DT_VARIANT]

  device='CPU'; TI in [DT_INT32]; T in [DT_VARIANT]

  device='CPU'; TI in [DT_UINT8]; T in [DT_VARIANT]

  device='CPU'; TI in [DT_INT64]; T in [DT_RESOURCE]

  device='CPU'; TI in [DT_INT32]; T in [DT_RESOURCE]

  device='CPU'; TI in [DT_UINT8]; T in [DT_RESOURCE]

  device='CPU'; TI in [DT_INT64]; T in [DT_STRING]

  device='CPU'; TI in [DT_INT32]; T in [DT_STRING]

  device='CPU'; TI in [DT_UINT8]; T in [DT_STRING]

  device='CPU'; TI in [DT_INT64]; T in [DT_BOOL]

  device='CPU'; TI in [DT_INT32]; T in [DT_BOOL]

  device='CPU'; TI in [DT_UINT8]; T in [DT_BOOL]

  device='CPU'; TI in [DT_INT64]; T in [DT_COMPLEX128]

  device='CPU'; TI in [DT_INT32]; T in [DT_COMPLEX128]

  device='CPU'; TI in [DT_UINT8]; T in [DT_COMPLEX128]

  device='CPU'; TI in [DT_INT64]; T in [DT_COMPLEX64]

  device='CPU'; TI in [DT_INT32]; T in [DT_COMPLEX64]

  device='CPU'; TI in [DT_UINT8]; T in [DT_COMPLEX64]

  device='CPU'; TI in [DT_INT64]; T in [DT_DOUBLE]

  device='CPU'; TI in [DT_INT32]; T in [DT_DOUBLE]

  device='CPU'; TI in [DT_UINT8]; T in [DT_DOUBLE]

  device='CPU'; TI in [DT_INT64]; T in [DT_FLOAT]

  device='CPU'; TI in [DT_INT32]; T in [DT_FLOAT]

  device='CPU'; TI in [DT_UINT8]; T in [DT_FLOAT]

  device='CPU'; TI in [DT_INT64]; T in [DT_BFLOAT16]

  device='CPU'; TI in [DT_INT32]; T in [DT_BFLOAT16]

  device='CPU'; TI in [DT_UINT8]; T in [DT_BFLOAT16]

  device='CPU'; TI in [DT_INT64]; T in [DT_HALF]

  device='CPU'; TI in [DT_INT32]; T in [DT_HALF]

  device='CPU'; TI in [DT_UINT8]; T in [DT_HALF]

  device='CPU'; TI in [DT_INT64]; T in [DT_INT8]

  device='CPU'; TI in [DT_INT32]; T in [DT_INT8]

  device='CPU'; TI in [DT_UINT8]; T in [DT_INT8]

  device='CPU'; TI in [DT_INT64]; T in [DT_UINT8]

  device='CPU'; TI in [DT_INT32]; T in [DT_UINT8]

  device='CPU'; TI in [DT_UINT8]; T in [DT_UINT8]

  device='CPU'; TI in [DT_INT64]; T in [DT_INT16]

  device='CPU'; TI in [DT_INT32]; T in [DT_INT16]

  device='CPU'; TI in [DT_UINT8]; T in [DT_INT16]

  device='CPU'; TI in [DT_INT64]; T in [DT_UINT16]

  device='CPU'; TI in [DT_INT32]; T in [DT_UINT16]

  device='CPU'; TI in [DT_UINT8]; T in [DT_UINT16]

  device='CPU'; TI in [DT_INT64]; T in [DT_INT32]

  device='CPU'; TI in [DT_INT32]; T in [DT_INT32]

  device='CPU'; TI in [DT_UINT8]; T in [DT_INT32]

  device='CPU'; TI in [DT_INT64]; T in [DT_INT64]

  device='CPU'; TI in [DT_INT32]; T in [DT_INT64]

  device='CPU'; TI in [DT_UINT8]; T in [DT_INT64]

 [Op:OneHot] name: one_hot/

Exception ignored in: >

Traceback (most recent call last):

  File "/home/kalarea/gym/gym/envs/classic_control/rendering.py", line 143, in __del__

  File "/home/kalarea/gym/gym/envs/classic_control/rendering.py", line 62, in close

  File "/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/pyglet/window/xlib/__init__.py", line 480, in close

  File "/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/pyglet/gl/xlib.py", line 345, in destroy

  File "/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/pyglet/gl/base.py", line 334, in destroy

  File "/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/pyglet/gl/xlib.py", line 335, in detach

  File "/home/kalarea/.conda/envs/py35/lib/python3.5/site-packages/pyglet/gl/lib.py", line 97, in errcheck

  File "", line 968, in _find_and_load

  File "", line 953, in _find_and_load_unlocked

  File "", line 887, in _find_spec

TypeError: 'NoneType' object is not iterable

Process finished with exit code 1