【发布时间】：2020-03-12 01:41:00
【问题描述】：
我定期导入 TensorFlow 概率，这是我的两层：
tfp.layers.DenseFlipout(512, activation=tf.nn.relu),
tfp.layers.DenseFlipout(10),
但是，当我运行此命令时，我不断收到此消息：
文件“dark.py”，第 10 行，在
import tensorflow_probability as tfp
File "C:\Python\Python37\lib\site-packages\tensorflow_probability\__init__.py", line 75, in <module>
from tensorflow_probability.python import * # pylint: disable=wildcard-import
File "C:\Python\Python37\lib\site-packages\tensorflow_probability\python\__init__.py", line 21, in <module>
from tensorflow_probability.python import bijectors
File "C:\Python\Python37\lib\site-packages\tensorflow_probability\python\bijectors\__init__.py", line 23, in <module>
from tensorflow_probability.python.bijectors.absolute_value import AbsoluteValue
File "C:\Python\Python37\lib\site-packages\tensorflow_probability\python\bijectors\absolute_value.py", line 23, in <module>
from tensorflow_probability.python.bijectors import bijector
File "C:\Python\Python37\lib\site-packages\tensorflow_probability\python\bijectors\bijector.py", line 213, in <module>
class Bijector(tf.Module):
File "C:\Python\Python37\lib\site-packages\tensorflow_probability\python\bijectors\bijector.py", line 570, in Bijector
_TF_MODULE_IGNORED_PROPERTIES = tf.Module._TF_MODULE_IGNORED_PROPERTIES.union(
AttributeError: type object 'Module' has no attribute '_TF_MODULE_IGNORED_PROPERTIES'
有什么办法可以解决这个问题吗？我在网上搜索过，但没有找到任何东西。
【问题讨论】：
-
是否可以用另一种方式编写bayes_prob？
-
到目前为止我有这个：
-
def bayes_prob(layer): with tf.compact.v1.name_scope("bayesian_prob", values=[layer]): model = tf.keras.Sequential([ tfp.layers.DenseFlipout(512 , 激活=tf.nn.relu), tfp.layers.DenseFlipout(10), ]) logits = model(features) neg_log_likelihood = tf.nn.softmax_cross_entropy_with_logits(labels=labels, logits=logits) kl = sum(model.losses ) loss = neg_log_likelihood + kl train_op = tf.train.AdamOptimizer().minimize(loss) 返回模型
-
检查this我认为是兼容性问题
标签： python neural-network deep-learning tensorflow2.0