no_grad

API属性：命令式编程模式(动态图)

paddle.fluid.dygraph. no_grad ( func=None ) [源代码]

创建一个上下文来禁用动态图梯度计算。在此模式下，每次计算的结果都将具有stop_gradient=True。

也可以用作一个装饰器（确保不要用括号来初始化）。

代码示例

import numpy as np
import paddle.fluid as fluid

# 用作生成器
data = np.array([[2, 3], [4, 5]]).astype('float32')
with fluid.dygraph.guard():
    l0 = fluid.Linear(2, 2)  # l0.weight.gradient() is None
    l1 = fluid.Linear(2, 2)
    with fluid.dygraph.no_grad():
        # l1.weight.stop_gradient is False
        tmp = l1.weight * 2  # tmp.stop_gradient is True
    x = fluid.dygraph.to_variable(data)
    y = l0(x) + tmp
    o = l1(y)
    o.backward()
    print(tmp.gradient() is None)  # True
    print(l0.weight.gradient() is None)  # False

# 用作装饰器
@fluid.dygraph.no_grad
def test_layer():
    with fluid.dygraph.guard():
        inp = np.ones([3, 1024], dtype='float32')
        t = fluid.dygraph.base.to_variable(inp)
        linear1 = fluid.Linear(1024, 4, bias_attr=False)
        linear2 = fluid.Linear(4, 4)
        ret = linear1(t)
        dy_ret = linear2(ret)

test_layer()