Pytorch production

vision
1

help me I’m having an error:
TypeError: forward() missing 1 required positional argument: ‘X’
this is my code
"import torch
import torchvision
import torch.nn as nn
import torchvision.transforms as transforms
from torch.autograd import Variable
from PIL import Image, ImageOps, ImageEnhance

class ConvolutionalNetwork(nn.Module):
def init(self):
super().init()
self.conv2 = nn.Conv2d(6, 16, kernel_size=3, stride=1,padding=1)
self.fc1 = nn.Linear(565616, 120)
#hidden layer de 120 y 84
self.fc2 = nn.Linear(120, 84)
self.fc3 = nn.Linear(84, 18)#este 18 son las 18 clases

def forward(self, X):
    X = F.relu(self.conv1(X))
    X = F.max_pool2d(X, 2, 2)#max pooling de 2x2
    X = F.relu(self.conv2(X))
    X = F.max_pool2d(X, 2, 2)
    X = X.view(-1, 56*56*16)
    X = F.relu(self.fc1(X))
    X = F.relu(self.fc2(X))
    X = self.fc3(X)
    return F.log_softmax(X, dim=1)

torch.manual_seed(101)
CNNmodel = ConvolutionalNetwork()
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(CNNmodel.parameters(), lr=0.001)

torch.manual_seed(0)
model=torch.load(‘quinua_pytorch.pt’)

model = CNNmodel()
model.load_state_dict(torch.load(‘quinua_pytorch.pt’)) #line to change
model.eval()

transform = transforms.Compose(
[transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]) # Same as for your validation data, e.g. Resize, ToTensor, Normalize, …

img = Image.open(’…/ejecutar/1.jpg’) # Load image as PIL.Image
x = transform(img) # Preprocess image
x = x.unsqueeze(0) # Add batch dimension

output = model(x) # Forward pass
pred = torch.argmax(output, 1) # Get predicted class if multi-class classification
print('Image predicted as ', pred)"

2

The posted code is not executable, as e.g. self.conv1 is missing.
Once this is fixed, one would need to remove other undefined classes until your code runs fine:

class ConvolutionalNetwork(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 6, 3, 1, 1)
        self.conv2 = nn.Conv2d(6, 16, kernel_size=3, stride=1,padding=1)
        self.fc1 = nn.Linear(56*56*16, 120)
        self.fc2 = nn.Linear(120, 84)
        self.fc3 = nn.Linear(84, 18)
    
    def forward(self, X):
        X = F.relu(self.conv1(X))
        X = F.max_pool2d(X, 2, 2)
        X = F.relu(self.conv2(X))
        X = F.max_pool2d(X, 2, 2)
        X = X.view(x.size(0), -1)
        X = F.relu(self.fc1(X))
        X = F.relu(self.fc2(X))
        X = self.fc3(X)
        return F.log_softmax(X, dim=1)

torch.manual_seed(101)
model = ConvolutionalNetwork()
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(CNNmodel.parameters(), lr=0.001)

x = torch.randn(1, 3, 224, 224)
output = model(x)
pred = torch.argmax(output, 1)

I’ve also changed the view operation, as it was failing since the input shape was undefined.

PS: you can post code snippets by wrapping them into three backticks ```, which makes debugging easier.
Also, once posted, copy/paste the code into a new script file, make sure it’s executable, and reproduces the error.

3

how would I adapt to predict an image I want to predict with the model quinua_pytorch.pt

4

I don’t know which model quinua_pytorch is, but to get a prediction for an input image you should reuse your previous image loading and processing code.
I.e. load the image, transform (and normalize) it to a tensor and get the prediction as seen in my code.

5

quinua_pytorch.pt is the training file after training the image dataset.
in your code you do not specify where I should load the file where it is already trained.

6

please support me I want to predict which class has a new image to the trained model, in your code it does not include the pytorch model load how it would be please