when I using MirroredStratage, the model didnt run the eval process.-.- #6
Description
Sorry to interrupt, but when I runing the code in MirroredStratage following the Tensorflow Guide, there's no eval process, the model is just traning all the time. .....
The code changes are as follows:
mirrored_strategy = tf.distribute.MirroredStrategy()
#config = tf.estimator.RunConfig(FLAGS.model_dir)
config = tf.estimator.RunConfig(model_dir=FLAGS.model_dir, train_distribute=mirrored_strategy)
what should I do about this? thank you so much!