Update '*var' according to the AdaMax algorithm.
tf.raw_ops.ApplyAdaMax(
var, m, v, beta1_power, lr, beta1, beta2, epsilon, grad, use_locking=False,
name=None
)
mt <- beta1 * m{t-1} + (1 - beta1) * g vt <- max(beta2 * v{t-1}, abs(g)) variable <- variable - learning_rate / (1 - beta1^t) * m_t / (v_t + epsilon)
Args | |
---|---|
var
|
A mutable Tensor . Must be one of the following types: float32 , float64 , int32 , uint8 , int16 , int8 , complex64 , int64 , qint8 , quint8 , qint32 , bfloat16 , uint16 , complex128 , half , uint32 , uint64 .
Should be from a Variable().
|
m
|
A mutable Tensor . Must have the same type as var .
Should be from a Variable().
|
v
|
A mutable Tensor . Must have the same type as var .
Should be from a Variable().
|
beta1_power
|
A Tensor . Must have the same type as var .
Must be a scalar.
|
lr
|
A Tensor . Must have the same type as var .
Scaling factor. Must be a scalar.
|
beta1
|
A Tensor . Must have the same type as var .
Momentum factor. Must be a scalar.
|
beta2
|
A Tensor . Must have the same type as var .
Momentum factor. Must be a scalar.
|
epsilon
|
A Tensor . Must have the same type as var .
Ridge term. Must be a scalar.
|
grad
|
A Tensor . Must have the same type as var . The gradient.
|
use_locking
|
An optional bool . Defaults to False .
If True , updating of the var, m, and v tensors will be protected
by a lock; otherwise the behavior is undefined, but may exhibit less
contention.
|
name
|
A name for the operation (optional). |
Returns | |
---|---|
A mutable Tensor . Has the same type as var .
|