AttributeError: module ‘torch.nn.parameter‘ has no attribute ‘UninitializedParameter‘

报错

Traceback (most recent call last):
  File "E:/Program Files/PyCharm 2019.2/PyG/test.py", line 63, in <module>
    model = GCN_NET(dataset.num_node_features, 16, dataset.num_classes).to(device)
  File "E:/Program Files/PyCharm 2019.2/PyG/test.py", line 45, in __init__
    self.conv1 = GCNConv(features, hidden)  # shape（输入的节点特征维度 * 中间隐藏层的维度）
  File "F:\Anaconda3\lib\site-packages\torch_geometric\nn\conv\gcn_conv.py", line 140, in __init__
    weight_initializer='glorot')
  File "F:\Anaconda3\lib\site-packages\torch_geometric\nn\dense\linear.py", line 65, in __init__
    self.reset_parameters()
  File "F:\Anaconda3\lib\site-packages\torch_geometric\nn\dense\linear.py", line 78, in reset_parameters
    if isinstance(self.weight, nn.parameter.UninitializedParameter):
AttributeError: module 'torch.nn.parameter' has no attribute 'UninitializedParameter'

解决

我当前的PyTorch版本为1.5.1，torch.nn.parameter确实没有UninitializedParameter这个属性，因此，需要对PyTorch进行升级。

不建议直接运行：

conda install pytorch torchvision torchaudio -c pytorch

建议在官网下载whl文件然后通过：

pip install path

其中path为文件路径。