本文整理匯總了Python中Math.abs方法的典型用法代碼示例。如果您正苦於以下問題：Python Math.abs方法的具體用法？Python Math.abs怎麽用？Python Math.abs使用的例子？那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類Math的用法示例。
在下文中一共展示了Math.abs方法的1個代碼示例，這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚，您的評價將有助於係統推薦出更棒的Python代碼示例。
示例1: proximal_descent
# 需要導入模塊: import Math [as 別名]
# 或者: from Math import abs [as 別名]
def proximal_descent(g, g_prime, h_prox, x0, iterations = 1000, gamma = 1.0, epsilon = 1e-4):
"""
minimizes a non-differentiable function f(x) = g(x) + h(x)
PARAMS
g: function
g(x), the differentiable part of f
g_prime: function
g'(x) aka the gradient of g
returns the direction of steepest increase along g
h_prox: function
h_prox(x, gamma) returns proximal operator of h at x using gamma as a distance weighting param
h_prox gives a new x' which is a tradeoff of reducing h and staying close to x
x0: vector
initial stariting point
iterations: self explanitory
gamma: step size
epsilon: self explanitory
RETURNS
x* = argmin_x { f(x) } if x* is reachable in the given num iterations. else None
"""
# initialize current guess at x0
xk = x0
gk = g(xk)
for _ in range(iterations):
xk_old = xk
# compute gradient for differentiable part of f
gk_gradient = g_prime(xk)
# take gradient step to reduce g(x)
xk_gradient = xk - gamma * gk_gradient
# proximal update to reduce h(x) but stay close to xk_gradient
xk = h_prox(xk_gradient, gamma)
if Math.abs(xk - xk_old) < epsilon:
return xk
return None