tf.contrib.mixed_precision.LossScaleOptimizer

View source on GitHub

An optimizer that applies loss scaling in backprop.

Inherits From: Optimizer

This class is useful for "mixed precision training" on GPUs (or other potential accelerators), an approach to improve compute throughput without compromising model quality.

The canonical way to perform mixed precision training is the following:

  • Model variables are kept in high precision (e.g. float32).
  • Computations are done in lower precision (e.g. float16), which enjoys performance speedup by virtue of hardware support. Variables are casted to lower precision before they're used.
  • Final gradients are casted back to high precision dtype, then used to update variables.

The side-effect of performing computation in lower precision, is that it comes with smaller numerical range. During backproping, small gradients might underflow in the reduced numerical range, causing a model to converge at suboptimal level.

To prevent underflow, this optimizer multiplies the loss by a factor before backprop starts. Consequently, the gradients are linearly scaled up by the same factor, thus not falling into the underflow zone. After that, to perserve the correctness of backprop, the gradients are down-scaled by the same factor, casted to the (higher) variable precision, then applied on the variables.

See Nvidia's manual on mixed precision training for more details.

To use loss scale optimizer, one only needs choose a loss scale strategy and wrap a regular optimizer. See examples below.

loss = loss_fn()
opt = tf.AdamOptimizer(learning_rate=...)

# Choose a loss scale manager which decides how to pick the right loss scale
# throughout the training process.
loss_scale_manager = tf.contrib.mixed_precision.FixedLossScaleManager(5000)

# Wraps the original optimizer in a LossScaleOptimizer.
loss_scale_optimizer =
    tf.contrib.mixed_precision.LossScaleOptimizer(opt, loss_scale_manager)

# Call minimize() on the loss scale optimizer.
train_op = loss_scale_optimizer.minimize(loss)

If gradients clipping is applied, one can call optimizer.compute_gradients() and optimizer.apply_gradients() separately.

Notice the following way of using LossScaleOptimizer is not intended. Always use loss_scale_optimizer.compute_gradients() to compute gradients instead of tf.gradients() if doing mixed precision training.

# The following is a wrong way to use LossScaleOptimizer along with
# tf.gradients().

# Always use loss_scale_optimizer.compute_gradients() to compute grads, or
# loss scale is not correctly applied.
grads = tf.gradients(loss, ...)

# Do some custom grad clipping.
grads = clip_grads(grads, ...)

loss_scale_optimizer.apply(grads_and_vars)

opt The actual optimizer that will be used to compute and apply the gradients. Must be an implementation of the tf.compat.v1.train.Optimizer interface.
loss_scale_manager A LossScaleManager object.

Methods

apply_gradients

View source

Apply gradients. See base class tf.compat.v1.train.Optimizer.

compute_gradients

View source

Compute gradients. See base class tf.compat.v1.train.Optimizer.

get_name

View source

get_slot

View source

Return a slot named name created for var by the Optimizer.

Some Optimizer subclasses use additional variables. For example Momentum and Adagrad use variables to accumulate updates. This method gives access to these Variable objects if for some reason you need them.

Use get_slot_names() to get the list of slot names created by the Optimizer.

Args
var A variable passed to minimize() or apply_gradients().
name A string.

Returns
The Variable for the slot if it was created, None otherwise.

get_slot_names

View source

Return a list of the names of slots created by the Optimizer.

See get_slot().

Returns
A list of strings.

minimize

View source

Add operations to minimize loss by updating var_list.

This method simply combines calls compute_gradients() and apply_gradients(). If you want to process the gradient before applying them call compute_gradients() and apply_gradients() explicitly instead of using this function.

Args
loss A Tensor containing the value to minimize.
global_step Optional Variable to increment by one after the variables have been updated.
var_list Optional list or tuple of Variable objects to update to minimize loss. Defaults to the list of variables collected in the graph under the key GraphKeys.TRAINABLE_VARIABLES.
gate_gradients How to gate the computation of gradients. Can be GATE_NONE, GATE_OP, or GATE_GRAPH.
aggregation_method Specifies the method used to combine gradient terms. Valid values are defined in the class AggregationMethod.
colocate_gradients_with_ops If True, try colocating gradients with the corresponding op.
name Optional name for the returned operation.
grad_loss Optional. A Tensor holding the gradient computed for loss.

Returns
An Operation that updates the variables in var_list. If global_step was not None, that operation also increments global_step.

Raises
ValueError If some of the variables are not Variable objects.

Eager Compatibility

When eager execution is enabled, loss should be a Python function that takes no arguments and computes the value to be minimized. Minimization (and gradient computation) is done with respect to the elements of var_list if not None, else with respect to any trainable variables created during the execution of the loss function. gate_gradients, aggregation_method, colocate_gradients_with_ops and grad_loss are ignored when eager execution is enabled.

variables

View source

A list of variables which encode the current state of Optimizer.

Includes slot variables and additional global variables created by the optimizer in the current default graph.

Returns
A list of variables.

Class Variables

  • GATE_GRAPH = 2
  • GATE_NONE = 0
  • GATE_OP = 1