Replaces the global generator with another
See Migration guide for more details.
tf.random.set_global_generator( generator )
This function creates a new Generator object (and the Variable object within), which does not work well with tf.function because (1) tf.function puts restrictions on Variable creation thus reset_global_generator can't be freely used inside tf.function; (2) redirecting a global variable to a new object is problematic with tf.function because the old object may be captured by a 'tf.function'ed function and still be used by it. A 'tf.function'ed function only keeps weak references to variables, so deleting a variable and then calling that function again may raise an error, as demonstrated by random_test.py/RandomTest.testResetGlobalGeneratorBadWithDefun .
| || the new |
© 2020 The TensorFlow Authors. All rights reserved.
Licensed under the Creative Commons Attribution License 3.0.
Code samples licensed under the Apache 2.0 License.