lstm_attention.py 训练问题

[sue2415535899]

您好，我在执行您的代码时运行到
out, attn, alpha = tf.split(lstm_net.outputs, [n_hidden, d_local, a_local ** 2], axis=2)
这一步报错：tensorflow.python.framework.errors_impl.InvalidArgumentError: Sum of output sizes must match the size of the original Tensor along the split dimension or the sum of the positive sizes must be less if it contains a -1 for 'split_1' (op: 'SplitV') with input shapes: [?,?,512], [3], [] and with computed input tensors: input[1] = <512 512 49>, input[2] = <2>.
我理解的是lstm_net.outputs 输出的是[?,?,512]维的tensor ,不能拆分为512，512，49维的out ,attn ,alpha。
我不知道是源代码问题，还是tensorflow 版本问题。
想问您原网络lstm_net输出的是[?,?,1073]还是[?,?,512]呢？

[qiongxiao]

我运行lstm_attention.py也遇到同样的问题，期待作者来解答。