squeezenet.py

from functools import partial
from typing import Any, Optional

import torch
import torch.nn as nn
import torch.nn.init as init

from ..transforms._presets import ImageClassification, InterpolationMode
from ..utils import _log_api_usage_once
from ._api import WeightsEnum, Weights
from ._meta import _IMAGENET_CATEGORIES
from ._utils import handle_legacy_interface, _ovewrite_named_param


__all__ = ["SqueezeNet", "SqueezeNet1_0_Weights", "SqueezeNet1_1_Weights", "squeezenet1_0", "squeezenet1_1"]


class Fire(nn.Module):
    def __init__(self, inplanes: int, squeeze_planes: int, expand1x1_planes: int, expand3x3_planes: int) -> None:
        super().__init__()
        self.inplanes = inplanes
        self.squeeze = nn.Conv2d(inplanes, squeeze_planes, kernel_size=1)
        self.squeeze_activation = nn.ReLU(inplace=True)
        self.expand1x1 = nn.Conv2d(squeeze_planes, expand1x1_planes, kernel_size=1)
        self.expand1x1_activation = nn.ReLU(inplace=True)
        self.expand3x3 = nn.Conv2d(squeeze_planes, expand3x3_planes, kernel_size=3, padding=1)
        self.expand3x3_activation = nn.ReLU(inplace=True)

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        x = self.squeeze_activation(self.squeeze(x))
        return torch.cat(
            [self.expand1x1_activation(self.expand1x1(x)), self.expand3x3_activation(self.expand3x3(x))], 1
        )


class SqueezeNet(nn.Module):
    def __init__(self, version: str = "1_0", num_classes: int = 1000, dropout: float = 0.5) -> None:
        super().__init__()
        _log_api_usage_once(self)
        self.num_classes = num_classes
        if version == "1_0":
            self.features = nn.Sequential(
                nn.Conv2d(3, 96, kernel_size=7, stride=2),
                nn.ReLU(inplace=True),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(96, 16, 64, 64),
                Fire(128, 16, 64, 64),
                Fire(128, 32, 128, 128),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(256, 32, 128, 128),
                Fire(256, 48, 192, 192),
                Fire(384, 48, 192, 192),
                Fire(384, 64, 256, 256),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(512, 64, 256, 256),
            )
        elif version == "1_1":
            self.features = nn.Sequential(
                nn.Conv2d(3, 64, kernel_size=3, stride=2),
                nn.ReLU(inplace=True),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(64, 16, 64, 64),
                Fire(128, 16, 64, 64),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(128, 32, 128, 128),
                Fire(256, 32, 128, 128),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(256, 48, 192, 192),
                Fire(384, 48, 192, 192),
                Fire(384, 64, 256, 256),
                Fire(512, 64, 256, 256),
            )
        else:
            # FIXME: Is this needed? SqueezeNet should only be called from the
            # FIXME: squeezenet1_x() functions
            # FIXME: This checking is not done for the other models
            raise ValueError(f"Unsupported SqueezeNet version {version}: 1_0 or 1_1 expected")

        # Final convolution is initialized differently from the rest
        final_conv = nn.Conv2d(512, self.num_classes, kernel_size=1)
        self.classifier = nn.Sequential(
            nn.Dropout(p=dropout), final_conv, nn.ReLU(inplace=True), nn.AdaptiveAvgPool2d((1, 1))
        )

        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                if m is final_conv:
                    init.normal_(m.weight, mean=0.0, std=0.01)
                else:
                    init.kaiming_uniform_(m.weight)
                if m.bias is not None:
                    init.constant_(m.bias, 0)

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        x = self.features(x)
        x = self.classifier(x)
        return torch.flatten(x, 1)


def _squeezenet(
    version: str,
    weights: Optional[WeightsEnum],
    progress: bool,
    **kwargs: Any,
) -> SqueezeNet:
    if weights is not None:
        _ovewrite_named_param(kwargs, "num_classes", len(weights.meta["categories"]))

    model = SqueezeNet(version, **kwargs)

    if weights is not None:
        model.load_state_dict(weights.get_state_dict(progress=progress))

    return model


_COMMON_META = {
    "task": "image_classification",
    "architecture": "SqueezeNet",
    "publication_year": 2016,
    "size": (224, 224),
    "categories": _IMAGENET_CATEGORIES,
    "interpolation": InterpolationMode.BILINEAR,
    "recipe": "https://github.com/pytorch/vision/pull/49#issuecomment-277560717",
}


class SqueezeNet1_0_Weights(WeightsEnum):
    IMAGENET1K_V1 = Weights(
        url="https://download.pytorch.org/models/squeezenet1_0-b66bff10.pth",
        transforms=partial(ImageClassification, crop_size=224),
        meta={
            **_COMMON_META,
            "min_size": (21, 21),
            "num_params": 1248424,
            "acc@1": 58.092,
            "acc@5": 80.420,
        },
    )
    DEFAULT = IMAGENET1K_V1


class SqueezeNet1_1_Weights(WeightsEnum):
    IMAGENET1K_V1 = Weights(
        url="https://download.pytorch.org/models/squeezenet1_1-b8a52dc0.pth",
        transforms=partial(ImageClassification, crop_size=224),
        meta={
            **_COMMON_META,
            "min_size": (17, 17),
            "num_params": 1235496,
            "acc@1": 58.178,
            "acc@5": 80.624,
        },
    )
    DEFAULT = IMAGENET1K_V1


@handle_legacy_interface(weights=("pretrained", SqueezeNet1_0_Weights.IMAGENET1K_V1))
def squeezenet1_0(
    *, weights: Optional[SqueezeNet1_0_Weights] = None, progress: bool = True, **kwargs: Any
) -> SqueezeNet:
    r"""SqueezeNet model architecture from the `"SqueezeNet: AlexNet-level
    accuracy with 50x fewer parameters and <0.5MB model size"
    <https://arxiv.org/abs/1602.07360>`_ paper.
    The required minimum input size of the model is 21x21.

    Args:
        weights (SqueezeNet1_0_Weights, optional): The pretrained weights for the model
        progress (bool): If True, displays a progress bar of the download to stderr
    """
    weights = SqueezeNet1_0_Weights.verify(weights)
    return _squeezenet("1_0", weights, progress, **kwargs)


@handle_legacy_interface(weights=("pretrained", SqueezeNet1_1_Weights.IMAGENET1K_V1))
def squeezenet1_1(
    *, weights: Optional[SqueezeNet1_1_Weights] = None, progress: bool = True, **kwargs: Any
) -> SqueezeNet:
    r"""SqueezeNet 1.1 model from the `official SqueezeNet repo
    <https://github.com/DeepScale/SqueezeNet/tree/master/SqueezeNet_v1.1>`_.
    SqueezeNet 1.1 has 2.4x less computation and slightly fewer parameters
    than SqueezeNet 1.0, without sacrificing accuracy.
    The required minimum input size of the model is 17x17.

    Args:
        weights (SqueezeNet1_1_Weights, optional): The pretrained weights for the model
        progress (bool): If True, displays a progress bar of the download to stderr
    """
    weights = SqueezeNet1_1_Weights.verify(weights)
    return _squeezenet("1_1", weights, progress, **kwargs)