Dropout¶

class Dropout(drop_prob=0.0, **kwargs)[源代码]¶

Randomly sets some elements of inputs to zeros with the probability \(drop\_prob\) during training. Commonly used in large networks for regularization and prevent overfitting, see Improving Neural Networks by Preventing Co-Adaptation of Feature Detectors<https://arxiv.org/abs/1207.0580>. Note that we perform dropout only during training, we also rescale(multiply) the output tensor by \(\frac{1}{1 - drop\_prob}\). During inference Dropout is equal to Identity.

参数: drop_prob (float) – The probability to drop (set to zero) each single element. Default: 0.0

Shape:

input: (*). Input can be of any shape.
output: (*). Output is of the same shape as input.

实际案例

>>> import numpy as np
>>> data = Tensor(np.ones(10000000, dtype=np.float32))
>>> out = F.nn.dropout(data, 1.0 / 3.0, training=True)
>>> assert not out.numpy().all()
>>> out = F.nn.dropout(data, 1.0 / 3.0, training=False)
>>> assert out.numpy().all()
>>> out.numpy()
array([1., 1., 1., ..., 1., 1., 1.], dtype=float32)

Linear

Embedding