等效于tensorflow中的np.add.at [英] Equivalent for np.add.at in tensorflow

问题描述

如何将np.add.at语句转换为tensorflow?

How do I convert a np.add.at statement into tensorflow?

np.add.at(dW, self.x.ravel(), dout.reshape(-1, self.D))

修改

self.dW.shape是(V，D)，self.D.shape是(N，D)，self.x.size是N

self.dW.shape is (V, D), self.D.shape is (N, D) and self.x.size is N

推荐答案

对于np.add.at，您可能希望查看

For np.add.at, you probably want to look at tf.SparseTensor, which represents a tensor by a list of values and a list of indices (which is more suitable for sparse data, hence the name).

以您的示例为例:

np.add.at(dW, self.x.ravel(), dout.reshape(-1, self.D))

应该是(假设dW，x和dout是张量):

that would be (assuming dW, x and dout are tensors):

tf.sparse_add(dW, tf.SparseTensor(x, tf.reshape(dout, [-1])))

这是假定x的形状为[n, nDims](即x是n个索引的列表"，每个索引的维度为nDims)，而dout的形状为[n].

This is assuming x is of shape [n, nDims] (i.e. x is a 'list' of n indices, each of dimension nDims), and dout has shape [n].

这篇关于等效于tensorflow中的np.add.at的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！