Skip to content

Instantly share code, notes, and snippets.

@sseveran
Created October 17, 2019 14:25
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save sseveran/88938abb055e9e2e7651a8b3fc3b13a9 to your computer and use it in GitHub Desktop.
Save sseveran/88938abb055e9e2e7651a8b3fc3b13a9 to your computer and use it in GitHub Desktop.
# Compute the minimize_op.
params = model.trainable_variables
grads = optimizer.get_gradients(total_loss, params)
grads_and_vars = list(zip(grads, params))
for gradient, variable in grads_and_vars:
var_name = variable.name.replace(":", "_")
tensorboard.summary.histogram(f"gradients/{var_name}", gradient)
tensorboard.summary.scalar(f"gradient_norm/{var_name}", tf.global_norm([gradient]))
minimize_op = optimizer.apply_gradients(grads_and_vars)
train_op = tf.group(minimize_op, update_ops)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment