Home > Mobile >  Pytorch ValueError: either size or scale_factor should be defined
Pytorch ValueError: either size or scale_factor should be defined

Time:12-13

Background Information I am trying to create a model (a beginner so please excuse my ignorance). The architecture I am trying to convert is given below as a link as well.

The Architecture to convert to a model

This is the code I came up with. I am using Binder to run the code.

import os
import torch
import torchvision
import tarfile
from torchvision.datasets.utils import download_url
from torch.utils.data import random_split
from torchsummary import summary
 
 
 
# Implementation of CNN/ConvNet Model
 
class build_unet(torch.nn.Module):
 
    def __init__(self):
        super(build_unet, self).__init__()
        
        keep_prob = 0.5 
        self.layer1 = torch.nn.Sequential(
            torch.nn.Conv2d(3, 32, kernel_size=3),
            torch.nn.ReLU(),
            torch.nn.MaxPool2d(kernel_size=2, padding=1))
       
        self.layer2 = torch.nn.Sequential(
            torch.nn.Conv2d(32, 64, kernel_size=3),
            torch.nn.ReLU(),
            torch.nn.MaxPool2d(kernel_size=2, padding=1))
       
        self.layer3 = torch.nn.Sequential(
            torch.nn.Conv2d(64, 128, kernel_size=3),
            torch.nn.ReLU(),
            torch.nn.MaxPool2d(kernel_size=2, padding=1))
 
        self.dense = torch.nn.Linear(64, 128, bias=True)
        torch.nn.init.xavier_uniform_(self.dense.weight)
        self.layer4 = torch.nn.Sequential(
            self.dense,
            torch.nn.ReLU(),
            torch.nn.Upsample()
        )
        
        self.layer5 = torch.nn.Sequential(
            torch.nn.Conv2d(128, 128, kernel_size=3),
            torch.nn.Sigmoid(),
            torch.nn.Upsample()
        )
        
        self.layer6 = torch.nn.Sequential(
            torch.nn.Conv2d(128, 64, kernel_size=3),
            torch.nn.Sigmoid(),
            torch.nn.Upsample()
        )
        
        self.layer7 = torch.nn.Sequential(
            torch.nn.Conv2d(64, 1, kernel_size=3),
            torch.nn.Sigmoid()
        )
        
    def forward(self, x):
        out = self.layer1(x)
        out = self.layer2(out)
        out = self.layer3(out)
        out = self.layer4(out)
        out = self.layer5(out)
        out = self.layer6(out)
        out = self.layer7(out)
     
        return out
 
if __name__ == "__main__":
    x = torch.randn((2, 3, 512, 512))
    f = build_unet()
    y = f(x)
    print(y.shape)

How would I resolve this error?

ERROR MESSAGE

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/tmp/ipykernel_36/1438699785.py in <module>
     87     x = torch.randn((2, 3, 512, 512))
     88     f = build_unet()
---> 89     y = f(x)
     90     print(y.shape)

/opt/conda/lib/python3.9/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
   1049         if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
   1050                 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1051             return forward_call(*input, **kwargs)
   1052         # Do not call functions when jit is used
   1053         full_backward_hooks, non_full_backward_hooks = [], []

/tmp/ipykernel_36/1438699785.py in forward(self, x)
     72         out = self.layer3(out)
     73 
---> 74         out = self.layer4(out)
     75         out = self.layer5(out)
     76         out = self.layer6(out)

/opt/conda/lib/python3.9/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
   1049         if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
   1050                 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1051             return forward_call(*input, **kwargs)
   1052         # Do not call functions when jit is used
   1053         full_backward_hooks, non_full_backward_hooks = [], []

/opt/conda/lib/python3.9/site-packages/torch/nn/modules/container.py in forward(self, input)
    137     def forward(self, input):
    138         for module in self:
--> 139             input = module(input)
    140         return input
    141 

/opt/conda/lib/python3.9/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
   1049         if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
   1050                 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1051             return forward_call(*input, **kwargs)
   1052         # Do not call functions when jit is used
   1053         full_backward_hooks, non_full_backward_hooks = [], []

/opt/conda/lib/python3.9/site-packages/torch/nn/modules/upsampling.py in forward(self, input)
    139 
    140     def forward(self, input: Tensor) -> Tensor:
--> 141         return F.interpolate(input, self.size, self.scale_factor, self.mode, self.align_corners)
    142 
    143     def extra_repr(self) -> str:

/opt/conda/lib/python3.9/site-packages/torch/nn/functional.py in interpolate(input, size, scale_factor, mode, align_corners, recompute_scale_factor)
   3647             scale_factors = [scale_factor for _ in range(dim)]
   3648     else:
-> 3649         raise ValueError("either size or scale_factor should be defined")
   3650 
   3651     if recompute_scale_factor is None:

ValueError: either size or scale_factor should be defined

CodePudding user response:

nn.Upsample() has following parameters: size, scale_factor, mode, align_corners. By default size=None, mode=nearest and align_corners=None.

torch.nn.Upsample(size=None, scale_factor=None, mode='nearest', align_corners=None)

When you set scale_factor=2 you will get following result:

import torch
import torch.nn as nn

class Net(torch.nn.Module):
 
    def __init__(self):
        super(Net, self).__init__()
        
        keep_prob = 0.5 
        self.layer1 = nn.Sequential(
            nn.Conv2d(3, 32, kernel_size=3),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, padding=1))
       
        self.layer2 = nn.Sequential(
            nn.Conv2d(32, 64, kernel_size=3),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, padding=1))
       
        self.layer3 = nn.Sequential(
            nn.Conv2d(64, 128, kernel_size=3),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, padding=1))
 
        self.dense = nn.Linear(64, 128, bias=True)
        nn.init.xavier_uniform_(self.dense.weight)
        self.layer4 = nn.Sequential(
            self.dense,
            nn.ReLU(),
            nn.Upsample(scale_factor=2)
        )
        
        self.layer5 = nn.Sequential(
            nn.Conv2d(128, 128, kernel_size=3),
            nn.Sigmoid(),
            nn.Upsample(scale_factor=2)
        )
        
        self.layer6 = nn.Sequential(
            nn.Conv2d(128, 64, kernel_size=3),
            nn.Sigmoid(),
            nn.Upsample(scale_factor=2)
        )
        
        self.layer7 = nn.Sequential(
            nn.Conv2d(64, 1, kernel_size=3),
            nn.Sigmoid()
        )
        
    def forward(self, x):
        out = self.layer1(x)
        out = self.layer2(out)
        out = self.layer3(out)
        out = self.layer4(out)
        out = self.layer5(out)
        out = self.layer6(out)
        out = self.layer7(out)
     
        return out
 
if __name__ == "__main__":
    x = torch.randn((2, 3, 512, 512))
    f = Net()
    y = f(x)
    print(y.shape)

Result:

torch.Size([2, 1, 498, 1010])
  • Related