1 问题描述

在使用PyTorch编程的时候，经常遇到一种报错就是：“RuntimeError: running_mean should contain *** elements not ***”；

这次我具体的报错信息是：

File "/home/***/software/python/anaconda/anaconda3/envs/conda-general/lib/python3.7/site-packages/torch/nn/modules/module.py", line 541, in __call__
    result = self.forward(*input, **kwargs)
File "/home/***/software/python/anaconda/anaconda3/envs/conda-general/lib/python3.7/site-packages/torch/nn/modules/container.py", line 92, in forward
    input = module(input)
File "/home/***/software/python/anaconda/anaconda3/envs/conda-general/lib/python3.7/site-packages/torch/nn/modules/module.py", line 541, in __call__
    result = self.forward(*input, **kwargs)
File "/home/***/software/python/anaconda/anaconda3/envs/conda-general/lib/python3.7/site-packages/torch/nn/modules/batchnorm.py", line 81, in forward
    exponential_average_factor, self.eps)
File "/home/***/software/python/anaconda/anaconda3/envs/conda-general/lib/python3.7/site-packages/torch/nn/functional.py", line 1670, in batch_norm
    training, momentum, eps, torch.backends.cudnn.enabled
RuntimeError: running_mean should contain 192 elements not 768

从最后一行的报错信息，可以看到：进行求均值元素的总数应该是192而不是768；

2 解决方案

我们可以继续看看上一条提示信息：“File "/home/***/software/python/anaconda/anaconda3/envs/conda-general/lib/python3.7/site-packages/torch/nn/functional.py", line 1670, in batch_norm”

有一个值得注意的信息是batch_norm，而我们的模型中也刚好使用了BN的操作，所以应该是BN的设置出现了问题，

我们回到代码定位的部分进行查看，需要查看的是BN初始化设置的代码，然后看到了下面的代码：

modules = [nn.Sequential(nn.Conv2d(in_channels, OUT_CHANNELS, 1, groups=1, bias=False),nn.BatchNorm2d(in_channels),nn.ReLU()),

我们可以看到，果然，BatchNorm2d的输入通道数与前一层Conv2d的输出通道数不一致，而这里的OUT_CHANNELS=192，in_channels=768，所以造成了这种维度的不一致，所以才会报错；

所以，我们需要根据自己模型的设计，将BN层与Conv层的输出维度保持一致。

PyTorch——解决报错“RuntimeError: running_mean should contain * elements not *”相关推荐

解决pytorch CrossEntropyLoss报错RuntimeError: 1D target tensor expected, multi-target not supported
解决方法 CrossEntropyLoss(预测值,label)需要的输入维度是: 有batch时,预测值维度为2,size为[ batch_size, n ]时,label的维度是1,size为[ ...
解决报错“RuntimeError - [Xcodeproj] Unknown object version.”
在 pod install 时提示错误: ### Error ``` RuntimeError - [Xcodeproj] Unknown object version. /Library/Ruby/ ...
pytorch并行报错RuntimeError: unable to write to file＜/torch_xxx＞
问题主要是使用multiprocessing在每一轮训练时并行随机产生数据时,job.get()数据超过了torch的暂存内存,参考了以下文章,加入代码或者降低了num_pool都没有解决我的问题 h ...
解决fitz模块报错RuntimeError: Directory ‘static/‘ does not exist
报错 fitz模块报错RuntimeError: Directory 'static/' does not exist 原因使用Python处理PDF文档时,需要使用fitz模块.由于Python ...
Pytorch_Geometric（PyG）使用DataLoader报错RuntimeError: Sizes of tensors must match except in dimension 0.
使用Pytorch_Geometric(PyG)时构建DataLoader,从DataLoader获取样本Batch时报错:RuntimeError: Sizes of tensors must ma ...
DDP报错::RuntimeError: Expected all tensors to be on the same device, but found at least two devic
DDP报错 RuntimeError: Expected all tensors to be on the same device, but found at least two devices, c ...
解决报错: MobaXterm X11 proxy: Unsupported authorisation protocol
此文首发于我的个人博客:解决报错 MobaXterm X11 proxy Unsupported authorisation protocol - zhang0peter的个人博客今天在使用Moba ...
Sublime Text 3便携版下载安装和常用插件安装--顺便解决报错An error occured installing和no packages available for install
文章目录 Sublime Text 3便携版下载安装和常用插件安装 1.Sublime Text 3便携版下载: 2.sublime 插件控制器(Package Control)安装 2.1离线安装P ...
Intel Realsense D435运行报错 RuntimeError: Camera not connected! dev.hardware_reset()函数需加睡眠sleep()
解决方案: 参考:Intel Realsense D435报错 RuntimeError: MFCreateDeviceSource(_device_attrs, &_source) retu ...

PyTorch——解决报错“RuntimeError: running_mean should contain * elements not *”

1 问题描述

2 解决方案

PyTorch——解决报错“RuntimeError: running_mean should contain * elements not *”相关推荐

最新文章

热门文章

PyTorch——解决报错“RuntimeError: running_mean should contain *** elements not ***”

1 问题描述

2 解决方案

PyTorch——解决报错“RuntimeError: running_mean should contain *** elements not ***”相关推荐

最新文章

热门文章

PyTorch——解决报错“RuntimeError: running_mean should contain * elements not *”

PyTorch——解决报错“RuntimeError: running_mean should contain * elements not *”相关推荐