AlexNet: Review and Implementation - deeplearning.vn

alexnet_lrn.py

"""
This is the implementation of AlexNet which is modified from [Jeicaoyu's AlexNet].

Note:
- The number of Conv2d filters now matches with the original paper.
- Use PyTorch's Local Response Normalization layer which is implemented in Jan 2018. [PR #4667]
- This is for educational purpose only. We don't have pretrained weights for this model.

References:

- Jeicaoyu's AlexNet Model: [jiecaoyu](https://github.com/jiecaoyu/pytorch_imagenet/blob/984a2a988ba17b37e1173dd2518fa0f4dc4a1879/networks/model_list/alexnet.py)
- PR #4667: https://github.com/pytorch/pytorch/pull/4667
"""
import torch.nn as nn


class AlexNet(nn.Module):

    def __init__(self, num_classes=1000):
        super(AlexNet, self).__init__()
        self.features = nn.Sequential(
            nn.Conv2d(3, 96, kernel_size=11, stride=4, padding=0),
            nn.ReLU(inplace=True),
            nn.LocalResponseNorm(size=5, alpha=0.0001, beta=0.75),
            nn.MaxPool2d(kernel_size=3, stride=2),
            nn.Conv2d(96, 256, kernel_size=5, padding=2, groups=2),
            nn.ReLU(inplace=True),
            nn.LocalResponseNorm(size=5, alpha=0.0001, beta=0.75),
            nn.MaxPool2d(kernel_size=3, stride=2),
            nn.Conv2d(256, 384, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
            nn.Conv2d(384, 384, kernel_size=3, padding=1, groups=2),
            nn.ReLU(inplace=True),
            nn.Conv2d(384, 256, kernel_size=3, padding=1, groups=2),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
        )
        self.classifier = nn.Sequential(
            nn.Linear(256 * 6 * 6, 4096),
            nn.ReLU(inplace=True),
            nn.Dropout(),
            nn.Linear(4096, 4096),
            nn.ReLU(inplace=True),
            nn.Dropout(),
            nn.Linear(4096, num_classes),
        )

    def forward(self, x):
        x = self.features(x)
        x = x.view(x.size(0), 256 * 6 * 6)
        x = self.classifier(x)
        return x

alexnet_torchvision.py

"""
This is AlexNet implementation from pytorch/torchvision.

Note:
- The number of nn.Conv2d doesn't match with the original paper.
- This model uses `nn.AdaptiveAvgPool2d` to allow the model to process images with arbitrary image size. [PR #746]
- This model doesn't use Local Response Normalization as described in the original paper.
    - This model is implemented in Jan 2017 with pretrained model.
    - PyTorch's Local Response Normalization layer is implemented in Jan 2018. [PR #4667]

References:
- Model: https://github.com/pytorch/vision/blob/ac2e995a4352267f65e7cc6d354bde683a4fb402/torchvision/models/alexnet.py
- PR #746: https://github.com/pytorch/vision/pull/746
- PR #4667: https://github.com/pytorch/pytorch/pull/4667
"""
import torch.nn as nn
from models.utils import load_state_dict_from_url

__all__ = ['AlexNet', 'alexnet']


model_urls = {
    'alexnet': 'https://download.pytorch.org/models/alexnet-owt-4df8aa71.pth',
}


class AlexNet(nn.Module):

    def __init__(self, num_classes=1000):
        super(AlexNet, self).__init__()
        self.features = nn.Sequential(
            nn.Conv2d(3, 64, kernel_size=11, stride=4, padding=2),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
            nn.Conv2d(64, 192, kernel_size=5, padding=2),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
            nn.Conv2d(192, 384, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
            nn.Conv2d(384, 256, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
            nn.Conv2d(256, 256, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
        )
        self.avgpool = nn.AdaptiveAvgPool2d((6, 6))
        self.classifier = nn.Sequential(
            nn.Dropout(),
            nn.Linear(256 * 6 * 6, 4096),
            nn.ReLU(inplace=True),
            nn.Dropout(),
            nn.Linear(4096, 4096),
            nn.ReLU(inplace=True),
            nn.Linear(4096, num_classes),
        )

    def forward(self, x):
        x = self.features(x)
        x = self.avgpool(x)
        x = x.view(x.size(0), 256 * 6 * 6)
        x = self.classifier(x)
        return x


def alexnet(pretrained=False, progress=True, **kwargs):
    r"""AlexNet model architecture from the
    `"One weird trick..." <https://arxiv.org/abs/1404.5997>`_ paper.

    Args:
        pretrained (bool): If True, returns a model pre-trained on ImageNet
        progress (bool): If True, displays a progress bar of the download to stderr
    """
    model = AlexNet(**kwargs)
    if pretrained:
        state_dict = load_state_dict_from_url(model_urls['alexnet'],
                                              progress=progress)
        model.load_state_dict(state_dict)
    return model