CODEWITHSUNDEEP
pythonmachine-learningpytorchmnist
mgus
mgus

Reputation: 808

PyTorch forward propagation returns different logits on same samples

Consider the following LeNet model for MNIST

import torch
from torch import nn
import torch.nn.functional as F

class LeNet(nn.Module):
    def __init__(self):
        super(LeNet, self).__init__()
        self.conv1 = nn.Conv2d(1, 20, 5, 1)
        self.conv2 = nn.Conv2d(20, 50, 5, 1)
        self.fc1 = nn.Linear(4*4*50, 500)
        self.fc2 = nn.Linear(500, 10)
        self.ceriation = nn.CrossEntropyLoss()
    def forward(self, x):
        x = self.conv1(x)
        x = F.max_pool2d(x, 2, 2)
        x = F.relu(x)
        x = self.conv2(x)
        x = F.max_pool2d(x, 2, 2)
        x = F.relu(x)
        x = x.view(-1, 4*4*50)
        x = self.fc1(x)
        x = self.fc2(x)
        return x

Now, I use this model to do a single forward step on a batch of samples like

network=LeNet()
optimizer = torch.optim.SGD(self.network.parameters(), lr=0.001, momentum=0.9)
device = torch.device("cpu")
network.to(device)
network.train()
optimizer.zero_grad()
# X_batch= ... some batch of 50 samples pulled from a train_loader defined as
# torch.manual_seed(42)
# training_set = datasets.MNIST('./mnist_data', train=True, download=False, 
#                               transform=transforms.Compose([
#                                   transforms.ToTensor(),
#                                   transforms.Normalize((0.1307,), (0.3081,))]))
# train_loader = torch.utils.data.DataLoader(training_set, 
#                                            batch_size=50, 
#                                            shuffle=False)
logits = network(X_batch)

Note that shuffle=False and download=False for the loader since the data set is already downloaded and I don't want to shuffle. My problem is that if I run this code twice I will get different values for logits and I don't understand why since everything else seems to be unchanged. For an extra check, I also extract X_batch to a numpy array and verify that the batch of samples is exactly the same as of previous execution. I do this check with numpy.array_equal() function.

I really can't figure out what I am missing here unless there are precision issues.

Upvotes: 1

Views: 340

Answers (1)

Alexander Pivovarov
Alexander Pivovarov

Reputation: 4990

The reason is because every time you run this code you call

network = LeNet()

and end up having different random initialization for the network's weights. If you set random seed before doing that, e.g. like this:

torch.manual_seed(42)
network = LeNet()

then you should get same results on first forward step given you use same data as input.

Upvotes: 2

Related Questions