RuntimeError: The size of tensor a must match the size of tensor b at non-singleton dimension 0 pytorch

[praveenkumarchandaliya]

def forward(self, input, hidden):
hx, cx = hidden
gates = self.conv_ih(input) + self.conv_hh(hx)

RuntimeError: The size of tensor a (32) must match the size of tensor b (18) at non-singleton dimension 0
i am using pytorch 0.3.1 version

[praveenkumarchandaliya]

i am using pytorch 0.3.1 version

[ssnl]

You used some op that expects two inputs to have same size, but they don't. This is not a bug in the framework, but in your code.

[praveenkumarchandaliya]

Same code running of 8581 image but when i have added any image then give this error

[ctivanovich]

I have this issue as well, using pytorch 1.3, but all of my images are square (600, 600, 3 color channels). I use transforms.Resize(224, Image.BICUBIC) when this error occurs. More specifically, RuntimeError: The size of tensor a (224) must match the size of tensor b (3) at non-singleton dimension 3 . It doesn't seem to be an issue of dimensionality on my end though, Image can display my images without trouble.

Also, just to confirm, i choose 224 for the size because I believe that is what resnet was trained on.

[ctivanovich]

Never mind, the error had nothing to do with Resize (I write this for anyone who encounters the same issue), it just seems that way because it's talking about dimensions. I was passing mean/std parameters to Normalize after the Resize step as an array rather than in a tuple: e.g. (m1, m2, m3).

[bhavika]

Another possible source of the issue could be that your C dimension from the tensor doesn't appear first. For example, if you have a tensor with shape (600, 600, 3) - the shape required for the transform may need to be (3, 600, 600).

You can achieve this using -

image = image.view(3, 600,600)

Now pass this image to your transform. Hope this helps!

[tehreemnaqvi]

Hi, I also got this error. How to fix this?

RuntimeError: The size of tensor a (8) must match the size of tensor b (4) at non-singleton dimension 3

[vickykhan89]

Hello all, good day!
I have two tensor a[100,8] and b[100,8], I need to change the shape/dimensions values in b[100,8] to b[8,100]. How can i do this need a piece of code. will be very thankful.

[rasmus-e-christensen]

To change a tensors shape, you need to transpose it using torch.transpose(your_tensor, 0, 1), which will transpose it on the 0 and 1 dimension.

[Akbarkhan87]

Error;
The size of tensor a (32) must match the size of tensor b (128) at non-singleton dimension 3

I want to change the network of vgg16 into vgg19 this type of error occurred, when I just train the model. Any one can help me in this regard @

[Akbarkhan87]

Model.py
What can i add and need your kind help to solve this issue as mentioned in the previous comment @bhavika
Error;
The size of tensor a (32) must match the size of tensor b (128) at non-singleton dimension 3

import torch.nn as nn
import torch
from torchvision import models
from utils import save_net,load_net

class CSRNet(nn.Module):
def init(self, load_weights=False):
super(CSRNet, self).init()
self.seen = 0
self.frontend_feat = [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M']
self.backend_feat = [512, 512, 512,256,128,64]
self.frontend = make_layers(self.frontend_feat)
self.backend = make_layers(self.backend_feat,in_channels = 512,dilation = True)
self.output_layer = nn.Conv2d(64, 1, kernel_size=1)
if not load_weights:
mod = models.vgg16(pretrained = True)
self.initialize_weights()
for i in range(len(self.frontend.state_dict().items())):
list(self.frontend.state_dict().items())[i][1].data[:] = list(mod.state_dict().items())[i][1].data[:]
def forward(self,x):
x = self.frontend(x)
x = self.backend(x)
x = self.output_layer(x)
return x
def initialize_weights(self):
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.normal(m.weight, std=0.01)
if m.bias is not None:
nn.init.constant(m.bias, 0)
elif isinstance(m, nn.BatchNorm2d):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)

def make_layers(cfg, in_channels = 3,batch_norm=False,dilation = False):
if dilation:
d_rate = 2
else:
d_rate = 1
layers = []
for v in cfg:
if v == 'M':
layers += [nn.MaxPool2d(kernel_size=2, stride=2)]
else:
conv2d = nn.Conv2d(in_channels, v, kernel_size=3, padding=d_rate,dilation = d_rate)
if batch_norm:
layers += [conv2d, nn.BatchNorm2d(v), nn.ReLU(inplace=True)]
else:
layers += [conv2d, nn.ReLU(inplace=True)]
in_channels = v
return nn.Sequential(*layers)

[eosayamen]

I have an error trying to use a github python script. Is there a way anyone can help me? I was trying to use a script that allows you to make an obj model from a photograph. Here is the link: GitHub - shunsukesaito/PIFu: This repository contains the code for the paper "PIFu: Pixel-Aligned Implicit Function for High-Resolution Clothed Human Digitization". I was able to install it using google collab. It worked with the original image, but when I changed the sample image I started seeing the error. I am not sure what the error means. The error states: The size of tensor a (245) must match the size of tensor b (244) at non-singleton dimension 2. Do you know if it means I need to resize my image? I think the error has something to do with the new image I added. Here is my link:
https://colab.research.google.com/drive/1-tAYm2kd5yNxGWZ-ooexSvd12V6f3ZQo?usp=sharing