tensorflow function笔记: dropout

最新推荐文章于 2022-12-31 11:48:12 发布

昵称123333

最新推荐文章于 2022-12-31 11:48:12 发布

阅读量339

点赞数

分类专栏： tensorflow 文章标签： tensorflow

本文链接：https://blog.csdn.net/rfliu666/article/details/61940006

版权

tensorflow 专栏收录该内容

4 篇文章 0 订阅

订阅专栏

Function

tf.nn.dropout(x, keep_prob, noise_shape=None, seed=None, name=None)

With probability keep_prob, outputs the input element scaled up by 1 / keep_prob, otherwise outputs 0. The scaling is so that the expected sum is unchanged.
输入元素有keep_prob 的几率按比例扩大1 / keep_prob 倍，否则输出为0. 比例为 1 / keep_prob 是为了使期望的总和不变。

By default, each element is kept or dropped independently. If noise_shape is specified, it must be broadcastable to the shape of x, and only dimensions with noise_shape[i] == shape(x)[i] will make independent decisions. For example, if shape(x) = [k, l, m, n] and noise_shape = [k, 1, 1, n], each batch and channel component will be kept independently and each row and column will be kept or not kept together.

默认每个元素被保留或者扔掉是独立的。如果指定 noise_shape, 只有 noise_shape[i] == shape(x)[i] 的维度是独立的。例如 shape(x) = [k, l, m, n] and noise_shape = [k, 1, 1, n]， batch 和 channel 的元素被保留的概率是独立， row 和 column 状态相同。
x_out = tf.nn.dropout(x, 0.5, noise_shape = [k, 1, 1, n])，则对于固定的k和n和任意的l和m， x_out[k, l, m, n]相等。(一组image的tensor的四个维度分别称作 [batch, row, column, channel ] 或者 [batch, height, width, channel] )