fused_dropout_add¶

参数¶

• x (Tensor): 输入张量。数据类型为 bfloat16, float16, float32 或 float64.

• y (Tensor): 输入张量，数据类型为 bfloat16, float16, float32 或 float64.

• p (float|int, 可选): 将单位设置为零的概率。默认值: 0.5。

• training (bool, 可选): 标记是否为训练阶段。默认值: True。

• mode (str, 可选): ['upscale_in_train'(默认) | 'downscale_in_infer'].

  1. upscale_in_train (默认), 在训练时放大输出

     • 训练: \(out = x \times \frac{mask}{(1.0 - dropout\_prob)} + y\)

     • 推理: \(out = x + y\)

  2. downscale_in_infer, 在推理时缩小输出

     • 训练: \(out = input \times mask + y\)

     • 推理: \(out = input \times (1.0 - dropout\_prob) + y\)

• name (str, 可选): 操作的名称, 默认值: None。有关更多信息, 请参阅 Name.

返回¶

• 融合 dropout 和 add 的张量具有与 x 相同的形状和数据类型。

代码示例¶

>>> import paddle
>>> from paddle.incubate.nn.functional import fused_dropout_add

>>> paddle.set_device('gpu')
>>> paddle.seed(2023)
>>> x = paddle.randn([4, 10], dtype="float32")
>>> y = paddle.randn([4, 10], dtype="float32")
>>> out = fused_dropout_add(x, y, p=0.5)
>>> print(out)
Tensor(shape=[4, 10], dtype=float32, place=Place(gpu:0), stop_gradient=True,
[[-0.49133155,  0.53819323, -2.58393312,  0.06336236, -1.09908366,
   0.22085167,  2.19751787,  0.05034769,  0.53417486,  0.84864247],
 [ 0.78248203, -1.59652555, -0.14399840, -0.77985179, -0.17006736,
  -0.30991879, -0.36593807, -0.51025450,  1.46401680,  0.61627960],
 [ 4.50472546, -0.48472026,  0.60729283,  0.33509624, -0.25593102,
  -1.45173049,  1.06727099,  0.00440830, -0.77340341,  0.67393088],
 [ 1.29453969,  0.07568165,  0.71947742, -0.71768606, -2.57172823,
   1.89179027,  3.26482797,  1.10493207, -1.04569530, -1.04862499]])