ValueError:形状必须是第 2 级，但对于“MatMul"是第 3 级 [英] ValueError: Shape must be rank 2 but is rank 3 for 'MatMul'

问题描述

我有以下 TensorFlow 代码:

I have the following TensorFlow code:

layer_1 = tf.add(tf.matmul(tf.cast(x, tf.float32), weights['h1']), biases['b1'])

但是抛出以下错误:

ValueError: Shape must be rank 2 but is rank 3 for 'MatMul' (op: 'MatMul') with input shapes: [?,5741,20000], [20000,128].

它说 x 的形状是 (?,5741,20000).如何将 x 的形状转换为 (5741, 20000)?

It says that x has the shape of (?,5741,20000). How could I transform the shape of x to (5741, 20000)?

先谢谢你！

推荐答案

我建议使用张量点积而不是简单的矩阵乘法以保持批量大小.这个答案比 @mrry

I would suggest to work with tensors dot product instead of simple matrix multiplication in order to keep the batch size. This is answer is more general than @mrry

layer_1 = tf.add(tf.tensordot(tf.cast(x, tf.float32), weights['h1'], [[2], [0]]), biases['b1'])

这篇关于ValueError:形状必须是第 2 级，但对于“MatMul"是第 3 级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！