# Copyright (C) 2020-2022, François-Guillaume Fernandez.
# This program is licensed under the Apache License 2.0.
# See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0> for full license details.
from typing import Any, Callable, Dict, List, Optional, Type, Union
from torch.nn import Module
from holocron.nn import PyConv2d
from ..presets import IMAGENETTE
from ..utils import conv_sequence, load_pretrained_params
from .resnet import ResNet, _ResBlock
__all__ = ["PyBottleneck", "pyconv_resnet50", "pyconvhg_resnet50"]
default_cfgs: Dict[str, Dict[str, Any]] = {
"pyconv_resnet50": {
**IMAGENETTE,
"input_shape": (3, 224, 224),
"url": None,
},
"pyconvhg_resnet50": {
**IMAGENETTE,
"input_shape": (3, 224, 224),
"url": None,
},
}
class PyBottleneck(_ResBlock):
expansion: int = 4
def __init__(
self,
inplanes: int,
planes: int,
stride: int = 1,
downsample: Optional[Module] = None,
groups: List[int] = [1],
base_width: int = 64,
dilation: int = 1,
act_layer: Optional[Module] = None,
norm_layer: Optional[Callable[[int], Module]] = None,
drop_layer: Optional[Callable[..., Module]] = None,
num_levels: int = 2,
**kwargs: Any,
) -> None:
width = int(planes * (base_width / 64.0)) * min(groups)
super().__init__(
[
*conv_sequence(
inplanes,
width,
act_layer,
norm_layer,
drop_layer,
kernel_size=1,
stride=1,
bias=(norm_layer is None),
**kwargs,
),
*conv_sequence(
width,
width,
act_layer,
norm_layer,
drop_layer,
conv_layer=PyConv2d,
kernel_size=3,
stride=stride,
padding=dilation,
groups=groups,
bias=(norm_layer is None),
dilation=dilation,
num_levels=num_levels,
**kwargs,
),
*conv_sequence(
width,
planes * self.expansion,
None,
norm_layer,
drop_layer,
kernel_size=1,
stride=1,
bias=(norm_layer is None),
**kwargs,
),
],
downsample,
act_layer,
)
class PyHGBottleneck(PyBottleneck):
expansion: int = 2
def _pyconvresnet(
arch: str,
pretrained: bool,
progress: bool,
block: Type[Union[PyBottleneck, PyHGBottleneck]],
num_blocks: List[int],
out_chans: List[int],
width_per_group: int,
groups: List[List[int]],
**kwargs: Any,
) -> ResNet:
# Build the model
model = ResNet(
block, # type: ignore[arg-type]
num_blocks,
out_chans,
stem_pool=False,
width_per_group=width_per_group,
block_args=[dict(num_levels=len(group), groups=group) for group in groups],
**kwargs,
)
model.default_cfg = default_cfgs[arch] # type: ignore[assignment]
# Load pretrained parameters
if pretrained:
load_pretrained_params(model, default_cfgs[arch]["url"], progress)
return model
[docs]
def pyconv_resnet50(pretrained: bool = False, progress: bool = True, **kwargs: Any) -> ResNet:
"""PyConvResNet-50 from `"Pyramidal Convolution: Rethinking Convolutional Neural Networks
for Visual Recognition" <https://arxiv.org/pdf/2006.11538.pdf>`_
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
Returns:
torch.nn.Module: classification model
"""
return _pyconvresnet(
"pyconv_resnet50",
pretrained,
progress,
PyBottleneck,
[3, 4, 6, 3],
[64, 128, 256, 512],
64,
[[1, 4, 8, 16], [1, 4, 8], [1, 4], [1]],
**kwargs,
)
[docs]
def pyconvhg_resnet50(pretrained: bool = False, progress: bool = True, **kwargs: Any) -> ResNet:
"""PyConvHGResNet-50 from `"Pyramidal Convolution: Rethinking Convolutional Neural Networks
for Visual Recognition" <https://arxiv.org/pdf/2006.11538.pdf>`_
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
Returns:
torch.nn.Module: classification model
"""
return _pyconvresnet(
"pyconvhg_resnet50",
pretrained,
progress,
PyHGBottleneck,
[3, 4, 6, 3],
[128, 256, 512, 1024],
2,
[[32, 32, 32, 32], [32, 64, 64], [32, 64], [32]],
**kwargs,
)