RT-DETR改进策略【卷积层】| 利用MobileNetv4中的UIB、ExtraDW优化ResNetLayer
一、本文介绍
本文记录的是利用
ExtraDW
优化
RT-DETR
中的
RepNCSPELAN4
，详细说明了优化原因，注意事项等。
ExtraDW
是
MobileNetv4
模型中提出的新模块，
允许以低成本增加网络深度和感受野，具有ConvNext和IB的组合优势
。可以在提高模型精度的同时降低一定量的模型参数。
二、UIB介绍
Universal Inverted Bottleneck（UIB）
通用反向瓶颈结构。
2.1 UIB结构设计
-
基于
MobileNetV4-
UIB建立在MobileNetV4之上，即采用深度可分离卷积和逐点扩展及投影的反向瓶颈结构。 -
在
反向瓶颈块（IB）中引入两个 可选的深度可分离卷积，一个在扩展层之前，另一个在扩展层和投影层之间。
-
-
UIB有四种可能的实例化形式：
- Inverted Bottleneck (IB) ：对扩展后的特征激活进行空间混合，以增加成本为代价提供更大的模型容量。
- ConvNext ：通过在扩展之前进行空间混合，使用更大的核尺寸实现更便宜的空间混合。
-
ExtraDW
：文中引入的新变体，允许以低成本增加网络深度和感受野，具有
ConvNext和IB的组合优势。 -
FFN
：由两个
1x1逐点卷积（PW）组成的栈，中间有激活和归一化层。
2.2 ExtraDW结构组成
结构组成 ：
-
在
IB块中加入两个可选的深度可分离卷积， 一个在扩展层之前，另一个在扩展层和投影层之间。
2.3 ExtraDW特点
-
灵活性 ：
- 在每个网络阶段，可以灵活地进行空间和通道混合的权衡调整，根据需要扩大感受野，并最大化计算利用率，增强模型对输入特征的感知能力。
-
效率提升 ：
- 提供了一种廉价增加网络深度和感受野的方式。相比其他结构，它在增加网络深度和感受野的同时，不会带来过高的计算成本。
- 在论文中，与其他注意力机制结合时，能有效提高模型的运算强度，减少内存访问需求，从而提高模型效率。
论文： http://arxiv.org/abs/2404.10518
源码： https://github.com/tensorflow/models/blob/master/official/vision/modeling/backbones/mobilenet.py
三、ExtraDW的实现代码
ExtraDW模块
的实现代码如下：
import torch
import torch.nn as nn
from typing import Optional
import torch.nn.functional as F
from ultralytics.utils.torch_utils import fuse_conv_and_bn
def make_divisible(
value: float,
divisor: int,
min_value: Optional[float] = None,
round_down_protect: bool = True,
) -> int:
"""
This function is copied from here
"https://github.com/tensorflow/models/blob/master/official/vision/modeling/layers/nn_layers.py"
This is to ensure that all layers have channels that are divisible by 8.
Args:
value: A `float` of original value.
divisor: An `int` of the divisor that need to be checked upon.
min_value: A `float` of minimum value threshold.
round_down_protect: A `bool` indicating whether round down more than 10%
will be allowed.
Returns:
The adjusted value in `int` that is divisible against divisor.
"""
if min_value is None:
min_value = divisor
new_value = max(min_value, int(value + divisor / 2) // divisor * divisor)
# Make sure that round down does not go down by more than 10%.
if round_down_protect and new_value < 0.9 * value:
new_value += divisor
return int(new_value)
def conv2d(in_channels, out_channels, kernel_size=3, stride=1, groups=1, bias=False, norm=True, act=True):
conv = nn.Sequential()
padding = (kernel_size - 1) // 2
conv.append(nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding, bias=bias, groups=groups))
if norm:
conv.append(nn.BatchNorm2d(out_channels))
if act:
conv.append(nn.ReLU6())
return conv
class UniversalInvertedBottleneckBlock(nn.Module):
def __init__(self, in_channels, out_channels, start_dw_kernel_size, middle_dw_kernel_size, middle_dw_downsample,
stride, expand_ratio):
"""An inverted bottleneck block with optional depthwises.
Referenced from here https://github.com/tensorflow/models/blob/master/official/vision/modeling/layers/nn_blocks.py
"""
super(UniversalInvertedBottleneckBlock, self).__init__()
# starting depthwise conv
self.start_dw_kernel_size = start_dw_kernel_size
if self.start_dw_kernel_size:
stride_ = stride if not middle_dw_downsample else 1
self._start_dw_ = conv2d(in_channels, in_channels, kernel_size=start_dw_kernel_size, stride=stride_, groups=in_channels, act=False)
# expansion with 1x1 convs
expand_filters = make_divisible(in_channels * expand_ratio, 8)
self._expand_conv = conv2d(in_channels, expand_filters, kernel_size=1)
# middle depthwise conv
self.middle_dw_kernel_size = middle_dw_kernel_size
if self.middle_dw_kernel_size:
stride_ = stride if middle_dw_downsample else 1
self._middle_dw = conv2d(expand_filters, expand_filters, kernel_size=middle_dw_kernel_size, stride=stride_, groups=expand_filters)
# projection with 1x1 convs
self._proj_conv = conv2d(expand_filters, out_channels, kernel_size=1, stride=1, act=False)
# expand depthwise conv (not used)
# _end_dw_kernel_size = 0
# self._end_dw = conv2d(out_channels, out_channels, kernel_size=_end_dw_kernel_size, stride=stride, groups=in_channels, act=False)
def forward(self, x):
if self.start_dw_kernel_size:
x = self._start_dw_(x)
# print("_start_dw_", x.shape)
x = self._expand_conv(x)
# print("_expand_conv", x.shape)
if self.middle_dw_kernel_size:
x = self._middle_dw(x)
# print("_middle_dw", x.shape)
x = self._proj_conv(x)
# print("_proj_conv", x.shape)
return x
def autopad(k, p=None, d=1): # kernel, padding, dilation
"""Pad to 'same' shape outputs."""
if d > 1:
k = d * (k - 1) + 1 if isinstance(k, int) else [d * (x - 1) + 1 for x in k] # actual kernel-size
if p is None:
p = k // 2 if isinstance(k, int) else [x // 2 for x in k] # auto-pad
return p
class Conv(nn.Module):
"""Standard convolution with args(ch_in, ch_out, kernel, stride, padding, groups, dilation, activation)."""
default_act = nn.SiLU() # default activation
def __init__(self, c1, c2, k=1, s=1, p=None, g=1, d=1, act=True):
"""Initialize Conv layer with given arguments including activation."""
super().__init__()
self.conv = nn.Conv2d(c1, c2, k, s, autopad(k, p, d), groups=g, dilation=d, bias=False)
self.bn = nn.BatchNorm2d(c2)
self.act = self.default_act if act is True else act if isinstance(act, nn.Module) else nn.Identity()
def forward(self, x):
"""Apply convolution, batch normalization and activation to input tensor."""
return self.act(self.bn(self.conv(x)))
def forward_fuse(self, x):
"""Perform transposed convolution of 2D data."""
return self.act(self.conv(x))
class ResNetBlock(nn.Module):
"""ResNet block with standard convolution layers."""
def __init__(self, c1, c2, s=1, e=4):
"""Initialize convolution with given parameters."""
super().__init__()
c3 = e * c2
self.cv1 = Conv(c1, c2, k=1, s=1, act=True)
self.cv2 = Conv(c2, c2, k=3, s=s, p=1, act=True)
self.cv3 = Conv(c2, c3, k=1, act=False)
self.cv4 = UniversalInvertedBottleneckBlock(c2, c2, 5, 3, True, 1, 4)
self.shortcut = nn.Sequential(Conv(c1, c3, k=1, s=s, act=False)) if s != 1 or c1 != c3 else nn.Identity()
def forward(self, x):
"""Forward pass through the ResNet block."""
return F.relu(self.cv3(self.cv4(self.cv2(self.cv1(x)))) + self.shortcut(x))
class ResNetLayer_UIB(nn.Module):
"""ResNet layer with multiple ResNet blocks."""
def __init__(self, c1, c2, s=1, is_first=False, n=1, e=4):
"""Initializes the ResNetLayer given arguments."""
super().__init__()
self.is_first = is_first
if self.is_first:
self.layer = nn.Sequential(
Conv(c1, c2, k=7, s=2, p=3, act=True), nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
)
else:
blocks = [ResNetBlock(c1, c2, s, e=e)]
blocks.extend([ResNetBlock(e * c2, c2, 1, e=e) for _ in range(n - 1)])
self.layer = nn.Sequential(*blocks)
def forward(self, x):
"""Forward pass through the ResNet layer."""
return self.layer(x)
四、添加步骤
4.1 改进点1
模块改进方法
1️⃣：直接加入
UniversalInvertedBottleneckBlock模块
。
UniversalInvertedBottleneckBlock模块
添加后如下：
注意❗：在
5.2和5.3小节
中需要声明的模块名称为：
UniversalInvertedBottleneckBlock
。
4.2 改进点2⭐
模块改进方法
2️⃣：基于
UniversalInvertedBottleneckBlock模块
的
ResNetLayer
。
第二种改进方法是对
RT-DETR
中的
ResNetLayer模块
进行改进。
UIB
中的
ExtraDW
模块与
ResNetLayer
结合后，可以为
RT-DETR
提供更丰富的特征表示，
更好地调整特征的空间分布和通道信息，使得模型能够更有效地聚焦于目标相关的特征，减少无关信息的干扰，进而提高检测精度。
改进代码如下：
首先添加
UniversalInvertedBottleneckBlock
模块改进
ResNetBlock
模块。
class ResNetBlock(nn.Module):
"""ResNet block with standard convolution layers."""
def __init__(self, c1, c2, s=1, e=4):
"""Initialize convolution with given parameters."""
super().__init__()
c3 = e * c2
self.cv1 = Conv(c1, c2, k=1, s=1, act=True)
self.cv2 = Conv(c2, c2, k=3, s=s, p=1, act=True)
self.cv3 = Conv(c2, c3, k=1, act=False)
self.cv4 = UniversalInvertedBottleneckBlock(c2, c2, 5, 3, True, 1, 4)
self.shortcut = nn.Sequential(Conv(c1, c3, k=1, s=s, act=False)) if s != 1 or c1 != c3 else nn.Identity()
def forward(self, x):
"""Forward pass through the ResNet block."""
return F.relu(self.cv3(self.cv4(self.cv2(self.cv1(x)))) + self.shortcut(x))
再添加如下代码将
ResNetLayer
重命名为
ResNetLayer_UIB
class ResNetLayer_UIB(nn.Module):
"""ResNet layer with multiple ResNet blocks."""
def __init__(self, c1, c2, s=1, is_first=False, n=1, e=4):
"""Initializes the ResNetLayer given arguments."""
super().__init__()
self.is_first = is_first
if self.is_first:
self.layer = nn.Sequential(
Conv(c1, c2, k=7, s=2, p=3, act=True), nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
)
else:
blocks = [ResNetBlock(c1, c2, s, e=e)]
blocks.extend([ResNetBlock(e * c2, c2, 1, e=e) for _ in range(n - 1)])
self.layer = nn.Sequential(*blocks)
def forward(self, x):
"""Forward pass through the ResNet layer."""
return self.layer(x)
注意❗：在
5.2和5.3小节
中需要声明的模块名称为：
ResNetLayer_UIB
。
五、添加步骤
5.1 修改一
① 在
ultralytics/nn/
目录下新建
AddModules
文件夹用于存放模块代码
② 在
AddModules
文件夹下新建
UIB.py
，将
第三节
中的代码粘贴到此处
5.2 修改二
在
AddModules
文件夹下新建
__init__.py
（已有则不用新建），在文件内导入模块：
from .UIB import *
5.3 修改三
在
ultralytics/nn/modules/tasks.py
文件中，需要在指定位置添加各模块类名称。
首先：导入模块
其次：在
parse_model函数
中注册
UniversalInvertedBottleneckBlock
和
ResNetLayer_UIB
模块
六、yaml模型文件
6.1 模型改进版本一
在代码配置完成后，配置模型的YAML文件。
此处以
ultralytics/cfg/models/rt-detr/rtdetr-l.yaml
为例，在同目录下创建一个用于自己数据集训练的模型文件
rtdetr-l-UIB.yaml
。
将
rtdetr-l.yaml
中的内容复制到
rtdetr-l-UIB.yaml
文件下，修改
nc
数量等于自己数据中目标的数量。
在骨干网络中添加
UniversalInvertedBottleneckBlock模块
，。
# Ultralytics YOLO 🚀, AGPL-3.0 license
# RT-DETR-l object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/rtdetr
# Parameters
nc: 1 # number of classes
scales: # model compound scaling constants, i.e. 'model=yolov8n-cls.yaml' will call yolov8-cls.yaml with scale 'n'
# [depth, width, max_channels]
l: [1.00, 1.00, 1024]
backbone:
# [from, repeats, module, args]
- [-1, 1, HGStem, [32, 48]] # 0-P2/4
- [-1, 6, UniversalInvertedBottleneckBlock, [48, 0, 3, True, 1, 2]] # stage 1
- [-1, 1, DWConv, [128, 3, 2, 1, False]] # 2-P3/8
- [-1, 6, UniversalInvertedBottleneckBlock, [128, 0, 3, True, 1, 2]] # stage 2
- [-1, 1, DWConv, [512, 3, 2, 1, False]] # 4-P4/16
- [-1, 6, UniversalInvertedBottleneckBlock, [512, 5, 3, True, 1, 4]] # cm, c2, k, light, shortcut
- [-1, 6, HGBlock, [192, 1024, 5, True, True]]
- [-1, 6, HGBlock, [192, 1024, 5, True, True]] # stage 3
- [-1, 1, DWConv, [1024, 3, 2, 1, False]] # 8-P5/32
- [-1, 6, UniversalInvertedBottleneckBlock, [1024, 5, 3, True, 1, 4]] # stage 4
head:
- [-1, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 10 input_proj.2
- [-1, 1, AIFI, [1024, 8]]
- [-1, 1, Conv, [256, 1, 1]] # 12, Y5, lateral_convs.0
- [-1, 1, nn.Upsample, [None, 2, "nearest"]]
- [7, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 14 input_proj.1
- [[-2, -1], 1, Concat, [1]]
- [-1, 3, RepC3, [256]] # 16, fpn_blocks.0
- [-1, 1, Conv, [256, 1, 1]] # 17, Y4, lateral_convs.1
- [-1, 1, nn.Upsample, [None, 2, "nearest"]]
- [3, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 19 input_proj.0
- [[-2, -1], 1, Concat, [1]] # cat backbone P4
- [-1, 3, RepC3, [256]] # X3 (21), fpn_blocks.1
- [-1, 1, Conv, [256, 3, 2]] # 22, downsample_convs.0
- [[-1, 17], 1, Concat, [1]] # cat Y4
- [-1, 3, RepC3, [256]] # F4 (24), pan_blocks.0
- [-1, 1, Conv, [256, 3, 2]] # 25, downsample_convs.1
- [[-1, 12], 1, Concat, [1]] # cat Y5
- [-1, 3, RepC3, [256]] # F5 (27), pan_blocks.1
- [[21, 24, 27], 1, RTDETRDecoder, [nc]] # Detect(P3, P4, P5)
6.2 模型改进版本二⭐
此处同样以
ultralytics/cfg/models/rt-detr/rtdetr-resnet50.yaml
为例，在同目录下创建一个用于自己数据集训练的模型文件
rtdetr-ResNetLayer_UIB.yaml
。
将
rtdetr-resnet50.yaml
中的内容复制到
rtdetr-ResNetLayer_UIB.yaml
文件下，修改
nc
数量等于自己数据中目标的数量。
📌 模型的修改方法是将
骨干网络
中的所有
ResNetLayer模块
替换成
ResNetLayer_UIB模块
。
# Parameters
nc: 1 # number of classes
scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n'
# [depth, width, max_channels]
m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs
backbone:
# [from, repeats, module, args]
- [-1, 1, Conv, [64, 3, 2]] # 0-P1/2
- [-1, 1, Conv, [128, 3, 2]] # 1-P2/4
- [-1, 3, C2f, [128, True]]
- [-1, 1, Conv, [256, 3, 2]] # 3-P3/8
- [-1, 6, C2f, [256, True]]
- [-1, 1, SCDown, [512, 3, 2]] # 5-P4/16
- [-1, 6, C2f, [512, True]]
- [-1, 1, SCDown, [1024, 3, 2]] # 7-P5/32
- [-1, 3, C2fCIB_UIB, [1024, True, True]]
- [-1, 1, SPPF, [1024, 5]] # 9
- [-1, 1, PSA, [1024]] # 10
# YOLOv8.0n head
head:
- [-1, 1, nn.Upsample, [None, 2, "nearest"]]
- [[-1, 6], 1, Concat, [1]] # cat backbone P4
- [-1, 3, C2f, [512]] # 13
- [-1, 1, nn.Upsample, [None, 2, "nearest"]]
- [[-1, 4], 1, Concat, [1]] # cat backbone P3
- [-1, 3, C2f, [256]] # 16 (P3/8-small)
- [-1, 1, Conv, [256, 3, 2]]
- [[-1, 13], 1, Concat, [1]] # cat head P4
- [-1, 3, C2f, [512]] # 19 (P4/16-medium)
- [-1, 1, SCDown, [512, 3, 2]]
- [[-1, 10], 1, Concat, [1]] # cat head P5
- [-1, 3, C2fCIB, [1024, True, True]] # 22 (P5/32-large)
- [[16, 19, 22], 1, v10Detect, [nc]] # Detect(P3, P4, P5)
七、成功运行结果
分别打印网络模型可以看到
UniversalInvertedBottleneckBlock
和
ResNetLayer_UIB
已经加入到模型中，并可以进行训练了。
rtdetr-l-UIB ：
rtdetr-l-UIB summary: 845 layers, 86,830,467 parameters, 86,830,467 gradients, 166.6 GFLOPs
from n params module arguments
0 -1 1 25248 ultralytics.nn.modules.block.HGStem [3, 32, 48]
1 -1 6 63360 ultralytics.nn.AddModules.UIB.UniversalInvertedBottleneckBlock[48, 48, 0, 3, True, 1, 2]
2 -1 1 3712 ultralytics.nn.modules.conv.DWConv [48, 128, 3, 2, 1, False]
3 -1 6 414720 ultralytics.nn.AddModules.UIB.UniversalInvertedBottleneckBlock[128, 128, 0, 3, True, 1, 2]
4 -1 1 5632 ultralytics.nn.modules.conv.DWConv [128, 512, 3, 2, 1, False]
5 -1 6 12831744 ultralytics.nn.AddModules.UIB.UniversalInvertedBottleneckBlock[512, 512, 5, 3, True, 1, 4]
6 -1 6 1695360 ultralytics.nn.modules.block.HGBlock [512, 192, 1024, 5, 6, True, True]
7 -1 6 2055808 ultralytics.nn.modules.block.HGBlock [1024, 192, 1024, 5, 6, True, True]
8 -1 1 11264 ultralytics.nn.modules.conv.DWConv [1024, 1024, 3, 2, 1, False]
9 -1 6 50829312 ultralytics.nn.AddModules.UIB.UniversalInvertedBottleneckBlock[1024, 1024, 5, 3, True, 1, 4]
10 -1 1 262656 ultralytics.nn.modules.conv.Conv [1024, 256, 1, 1, None, 1, 1, False]
11 -1 1 789760 ultralytics.nn.modules.transformer.AIFI [256, 1024, 8]
12 -1 1 66048 ultralytics.nn.modules.conv.Conv [256, 256, 1, 1]
13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
14 7 1 262656 ultralytics.nn.modules.conv.Conv [1024, 256, 1, 1, None, 1, 1, False]
15 [-2, -1] 1 0 ultralytics.nn.modules.conv.Concat [1]
16 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
17 -1 1 66048 ultralytics.nn.modules.conv.Conv [256, 256, 1, 1]
18 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
19 3 1 33280 ultralytics.nn.modules.conv.Conv [128, 256, 1, 1, None, 1, 1, False]
20 [-2, -1] 1 0 ultralytics.nn.modules.conv.Concat [1]
21 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
22 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2]
23 [-1, 17] 1 0 ultralytics.nn.modules.conv.Concat [1]
24 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
25 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2]
26 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1]
27 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
28 [21, 24, 27] 1 7303907 ultralytics.nn.modules.head.RTDETRDecoder [1, [256, 256, 256]]
rtdetr-l-UIB summary: 845 layers, 86,830,467 parameters, 86,830,467 gradients, 166.6 GFLOPs
rtdetr-ResNetLayer_UIB ：
rtdetr-ResNetLayer_UIB summary: 833 layers, 53,128,419 parameters, 53,128,419 gradients, 159.6 GFLOPs
from n params module arguments
0 -1 1 9536 ultralytics.nn.AddModules.UIB.ResNetLayer_UIB[3, 64, 1, True, 1]
1 -1 1 329664 ultralytics.nn.AddModules.UIB.ResNetLayer_UIB[64, 64, 1, False, 3]
2 -1 1 1785344 ultralytics.nn.AddModules.UIB.ResNetLayer_UIB[256, 128, 2, False, 4]
3 -1 1 10368512 ultralytics.nn.AddModules.UIB.ResNetLayer_UIB[512, 256, 2, False, 6]
4 -1 1 21380608 ultralytics.nn.AddModules.UIB.ResNetLayer_UIB[1024, 512, 2, False, 3]
5 -1 1 524800 ultralytics.nn.modules.conv.Conv [2048, 256, 1, 1, None, 1, 1, False]
6 -1 1 789760 ultralytics.nn.modules.transformer.AIFI [256, 1024, 8]
7 -1 1 66048 ultralytics.nn.modules.conv.Conv [256, 256, 1, 1]
8 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
9 3 1 262656 ultralytics.nn.modules.conv.Conv [1024, 256, 1, 1, None, 1, 1, False]
10 [-2, -1] 1 0 ultralytics.nn.modules.conv.Concat [1]
11 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
12 -1 1 66048 ultralytics.nn.modules.conv.Conv [256, 256, 1, 1]
13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
14 2 1 131584 ultralytics.nn.modules.conv.Conv [512, 256, 1, 1, None, 1, 1, False]
15 [-2, -1] 1 0 ultralytics.nn.modules.conv.Concat [1]
16 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
17 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2]
18 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1]
19 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
20 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2]
21 [-1, 7] 1 0 ultralytics.nn.modules.conv.Concat [1]
22 -1 3 2232320 ultralytics.nn.modules.block.RepC3 [512, 256, 3]
23 [16, 19, 22] 1 7303907 ultralytics.nn.modules.head.RTDETRDecoder [1, [256, 256, 256]]
rtdetr-ResNetLayer_UIB summary: 833 layers, 53,128,419 parameters, 53,128,419 gradients, 159.6 GFLOPs