Python 如何修复在GRU'中使用DataParallel时的错误；s网络

我正试图像文档中解释的那样将dataparallel应用到GRU的网络中，我不断地得到相同的错误

"""Defines the neural network, losss function and metrics"""

import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F


class Net(nn.Module):
    def __init__(self, params, anchor_is_phrase):
        """
        Simple LSTM, used to generate the LSTM for both the word and video
        embeddings. 
        Args:
            params: (Params) contains vocab_size, embedding_dim, lstm_hidden_dim
            is_phrase: is word lstm or the vid lstm
        """
        super(Net, self).__init__()
        if anchor_is_phrase:
            self.lstm = nn.DataParallel(nn.GRU(params.word_embedding_dim, params.hidden_dim, 1)).cuda()#, batch_first=True)
        else:
            self.lstm = nn.DataParallel(nn.GRU(params.vid_embedding_dim, params.hidden_dim, 1)).cuda() #, batch_first=True)

    def forward(self, s, anchor_is_phrase = False):
        """
        Forward prop. 
        """
        s, _ = self.lstm(s)
        s.data.contiguous()
        return s

错误发生在前一代码中的第s行=self.lstm处：

 here:  s, _ = self.lstm(s)
        s.data.contiguous()
        return s

我收到以下错误消息：

    s, _ = self.lstm(s)
  File "/home/pavelameen/miniconda3/envs/TD2/lib/python3.6/site-packages/torch/nn/modules/module.py", line 493, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/pavelameen/miniconda3/envs/TD2/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 152, in forward
    outputs = self.parallel_apply(replicas, inputs, kwargs)
  File "/home/pavelameen/miniconda3/envs/TD2/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 162, in parallel_apply
    return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
  File "/home/pavelameen/miniconda3/envs/TD2/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 83, in parallel_apply
    raise output
  File "/home/pavelameen/miniconda3/envs/TD2/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 59, in _worker
    output = module(*input, **kwargs)
  File "/home/pavelameen/miniconda3/envs/TD2/lib/python3.6/site-packages/torch/nn/modules/module.py", line 493, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/pavelameen/miniconda3/envs/TD2/lib/python3.6/site-packages/torch/nn/modules/rnn.py", line 193, in forward
    max_batch_size = input.size(0) if self.batch_first else input.size(1)
AttributeError: 'tuple' object has no attribute 'size'

---

有趣的是，我试图在第27行中输出s的类型，我得到了PackedSequence，为什么它在lstm正向方法中转换为元组？

nn.GRU

（第181行）a或tesnor作为输入。如错误中所述，您正在传递一个元组对象

s

intead.

nn.GRU

（第181行）或tesnor作为输入。如错误中所述，您正在传递一个元组对象

s

intead。

在传递到nn.GRU之前，我打印了s的类型，并获得了PackedSequence！！！这很奇怪，因为您的错误来自一行，该行只有在输入不是

PackedSequence

时才会执行。参见上述链接中的第181至191行。或者你可以查看你正在使用的pytorch的相应版本。在传递到nn.GRU之前，我打印了s的类型，我得到了PackedSequence！！！这很奇怪，因为您的错误来自一行，该行只有在输入不是

PackedSequence

时才会执行。参见上述链接中的第181至191行。或者，您可以查看正在使用的pytorch的相应版本。