torch.compile (inductor)

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

class Model(nn.Module):
    def __init__(self, num_classes):
        """
        LeNet-5 architecture implementation in PyTorch.

        :param num_classes: The number of output classes.
        """
        super(Model, self).__init__()
        
        # Convolutional layers
        self.conv1 = nn.Conv2d(in_channels=1, out_channels=6, kernel_size=5, stride=1)
        self.conv2 = nn.Conv2d(in_channels=6, out_channels=16, kernel_size=5, stride=1)
        
        # Fully connected layers
        self.fc1 = nn.Linear(in_features=16*5*5, out_features=120)
        self.fc2 = nn.Linear(in_features=120, out_features=84)
        self.fc3 = nn.Linear(in_features=84, out_features=num_classes)
    
    def forward(self, x):
        """
        Forward pass of the LeNet-5 model.

        :param x: The input tensor, shape (batch_size, 1, 32, 32)
        :return: The output tensor, shape (batch_size, num_classes)
        """
        # First convolutional layer with ReLU activation and max pooling
        x = F.relu(self.conv1(x))
        x = F.max_pool2d(x, kernel_size=2, stride=2)
        
        # Second convolutional layer with ReLU activation and max pooling
        x = F.relu(self.conv2(x))
        x = F.max_pool2d(x, kernel_size=2, stride=2)
        
        # Flatten the output for the fully connected layers
        x = x.view(-1, 16*5*5)
        
        # First fully connected layer with ReLU activation
        x = F.relu(self.fc1(x))
        
        # Second fully connected layer with ReLU activation
        x = F.relu(self.fc2(x))
        
        # Final fully connected layer
        x = self.fc3(x)
        
        return x

# Test code for the LeNet-5 model (larger batch & image)
batch_size = 4096
num_classes = 20

def get_inputs():
    return [torch.rand(batch_size, 1, 32, 32)]

def get_init_inputs():
    return [num_classes]