# Copyright (c) 2021 zfit
import tensorflow as tf
from .base_tf import WrapOptimizer
[docs]class Adam(WrapOptimizer):
# todo see why sphinx does not correctly link to WrapOptimizer for api
_DEFAULT_name = 'Adam'
def __init__(self,
tol=None,
learning_rate=0.2,
beta1=0.9,
beta2=0.999,
epsilon=1e-08,
name='Adam', **kwargs):
# todo write documentation for api.
optimizer = tf.keras.optimizers.Adam(learning_rate=learning_rate,
beta_1=beta1, beta_2=beta2,
epsilon=epsilon,
name=name)
super().__init__(optimizer=optimizer, tol=tol, maxiter=None, **kwargs)