定义PyTorch模型时候遇到问题
1
2
3
4
5
6
| TypeError: new() received an invalid combination of arguments - got (float, int, int, int), but expected one of:
* (torch.device device)
* (torch.Storage storage)
* (Tensor other)
* (tuple of ints size, torch.device device)
* (object data, torch.device device)
|
出错代码为
1
| self.conv1 = nn.Conv2d(input_channels, output_channels/4, 1, 1, bias = False)
|
由于环境是python3,整除会出现float型，造成参数类型错误，这里很有可能就是output_channels/4出了问题。
在python3中，1024/4=256.0
1
2
3
4
5
| Python 3.7.4 (default, Aug 13 2019, 20:35:49)
[GCC 7.3.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> 1024/4
256.0
|
而在python2中，1024/4=256
1
2
3
4
5
| Python 2.7.18rc1 (default, Apr 7 2020, 12:05:55)
[GCC 9.3.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> 1024/4
256
|
将/改为//，表示强制转换为int型，问题解决。
1
| self.conv1 = nn.Conv2d(input_channels, output_channels//4, 1, 1, bias = False)
|