LostTech.TensorFlow : API Documentation

Type I_TransformGradients

Namespace tensorflow_estimator.contrib.estimator

Interfaces IOptimizer

Public instance methods

object apply_gradients(object grads_and_vars, object global_step, object name)

object compute_gradients(IDictionary<string, object> kwargs, Object[] args)

Compute gradients of "loss" for the variables in "var_list".

This simply wraps the compute_gradients() from the real optimizer. The gradients will be aggregated in the apply_gradients() so that user can modify the gradients like clipping with per replica global norm if needed. The global norm with aggregated gradients can be bad as one replica's huge gradients can hurt the gradients from other replicas.
IDictionary<string, object> kwargs
Keyword arguments for compute_gradients().
Object[] args
Arguments for compute_gradients().
A list of (gradient, variable) pairs.