What does global_step mean in Tensorflow ?

What does global_step mean in Tensorflow ?

Asked on December 15, 2018 in Tensorflow.
Add Comment


  • 2 Answer(s)

           The global_step means the number of batches seen by the graph. Each time a batch is given, the weights are updated toward the path that minimizes the loss. global_step just monitors the number of batches. When it is passed in the minimize() argument list, the variable is increased by one and also see here optimizer.minimize() The initial value is 0 for the global step in this context.

           Get the global_step value by using tf.train.global_step(). The advanced methods are tf.train.get_global_step or tf.train.get_or_create_global_step.

    Answered on December 15, 2018.
    Add Comment

          In Tensorflow, the global_step variable holds the total number of steps during training over the tasks.

          Each step index will occur only on a single task. For Example, the loss and accuracy would be plotted against global_step on Tensorboard.

    Answered on December 15, 2018.
    Add Comment


  • Your Answer

    By posting your answer, you agree to the privacy policy and terms of service.