def testDecay(self):
num_training_steps = 1000
initial_lr = 1.0
for step in range(0, 1500, 250):
decayed_lr = learning_rate_decay_v2.cosine_decay(initial_lr, step,
num_training_steps)
expected = self.np_cosine_decay(step, num_training_steps)
self.assertAllClose(self.evaluate(decayed_lr()), expected, 1e-6)
def testDecay(self):
num_training_steps = 1000
initial_lr = 1.0
for step in range(0, 1500, 250):
decayed_lr = learning_rate_decay_v2.cosine_decay(initial_lr, step,
num_training_steps)
expected = self.np_cosine_decay(step, num_training_steps)
self.assertAllClose(self.evaluate(decayed_lr()), expected, 1e-6)
def cosine_decay(learning_rate, global_step, decay_steps, alpha=0.0, name=None):
"""Applies cosine decay to the learning rate.
See [Loshchilov & Hutter, ICLR2016], SGDR: Stochastic Gradient Descent
with Warm Restarts. https://arxiv.org/abs/1608.03983
When training a model, it is often recommended to lower the learning rate as
the training progresses. This function applies a cosine decay function
to a provided initial learning rate. It requires a `global_step` value to
compute the decayed learning rate. You can just pass a TensorFlow variable
that you increment at each training step.
The function returns the decayed learning rate. It is computed as:
```python
global_step = min(global_step, decay_steps)
cosine_decay = 0.5 * (1 + cos(pi * global_step / decay_steps))
decayed = (1 - alpha) * cosine_decay + alpha
decayed_learning_rate = learning_rate * decayed
```
Example usage:
```python
decay_steps = 1000
lr_decayed = cosine_decay(learning_rate, global_step, decay_steps)
```
Args:
learning_rate: A scalar `float32` or `float64` Tensor or a Python number.
The initial learning rate.
global_step: A scalar `int32` or `int64` `Tensor` or a Python number.
Global step to use for the decay computation.
decay_steps: A scalar `int32` or `int64` `Tensor` or a Python number.
Number of steps to decay over.
alpha: A scalar `float32` or `float64` Tensor or a Python number.
Minimum learning rate value as a fraction of learning_rate.
name: String. Optional name of the operation. Defaults to 'CosineDecay'.
Returns:
A scalar `Tensor` of the same type as `learning_rate`. The decayed
learning rate.
Raises:
ValueError: if `global_step` is not supplied.
@compatibility(eager)
When eager execution is enabled, this function returns a function which in
turn returns the decayed learning rate Tensor. This can be useful for changing
the learning rate value across different invocations of optimizer functions.
@end_compatibility
"""
decayed_lr = learning_rate_decay_v2.cosine_decay(
learning_rate, global_step, decay_steps, alpha=alpha, name=name)
if not context.executing_eagerly():
decayed_lr = decayed_lr()
return decayed_lr
def cosine_decay(learning_rate,
global_step,
decay_steps,
alpha=0.0,
name=None):
"""Applies cosine decay to the learning rate.
See [Loshchilov & Hutter, ICLR2016], SGDR: Stochastic Gradient Descent
with Warm Restarts. https://arxiv.org/abs/1608.03983
When training a model, it is often recommended to lower the learning rate as
the training progresses. This function applies a cosine decay function
to a provided initial learning rate. It requires a `global_step` value to
compute the decayed learning rate. You can just pass a TensorFlow variable
that you increment at each training step.
The function returns the decayed learning rate. It is computed as:
```python
global_step = min(global_step, decay_steps)
cosine_decay = 0.5 * (1 + cos(pi * global_step / decay_steps))
decayed = (1 - alpha) * cosine_decay + alpha
decayed_learning_rate = learning_rate * decayed
```
Example usage:
```python
decay_steps = 1000
lr_decayed = cosine_decay(learning_rate, global_step, decay_steps)
```
Args:
learning_rate: A scalar `float32` or `float64` Tensor or a Python number.
The initial learning rate.
global_step: A scalar `int32` or `int64` `Tensor` or a Python number.
Global step to use for the decay computation.
decay_steps: A scalar `int32` or `int64` `Tensor` or a Python number.
Number of steps to decay over.
alpha: A scalar `float32` or `float64` Tensor or a Python number.
Minimum learning rate value as a fraction of learning_rate.
name: String. Optional name of the operation. Defaults to 'CosineDecay'.
Returns:
A scalar `Tensor` of the same type as `learning_rate`. The decayed
learning rate.
Raises:
ValueError: if `global_step` is not supplied.
@compatibility(eager)
When eager execution is enabled, this function returns a function which in
turn returns the decayed learning rate Tensor. This can be useful for changing
the learning rate value across different invocations of optimizer functions.
@end_compatibility
"""
decayed_lr = learning_rate_decay_v2.cosine_decay(learning_rate,
global_step,
decay_steps,
alpha=alpha,
name=name)
if not context.executing_eagerly():
decayed_lr = decayed_lr()
return decayed_lr