当前位置:网站首页>【MnasNet】《MnasNet:Platform-Aware Neural Architecture Search for Mobile》
【MnasNet】《MnasNet:Platform-Aware Neural Architecture Search for Mobile》
2022-07-02 06:26:00 【bryant_meng】


CVPR-2019
文章目录
1 Background and Motivation
作者旨在设计一个新的 resource-constrained mobile model 让其在 resource-constrained platforms 跑的更加欢快
2 Related Work
现有网络的基础上压缩:量化,pruning ,NetAdapt 等,do not focus on learning novel compositions of CNN operations
hand-crafted 设计,usually take significant human efforts
NAS,基于各种 learning algorithms,例如 reinforcement learning / evolutionary search / differentiable search
3 Advantages / Contributions
NAS 出 MnasNet,两个主要创新点
incorporate model latency into the main objective so that the search can identify a model that achieves a good trade-off between accuracy and latency(不单单是 ACC)
a novel factorized hierarchical search space that encourages layer diversity throughout the network.(不像 NasNet 那样是 cell 级别的,而是 block 级别的)
achieve new state-of-the-art results on both ImageNet classification and COCO object detection under typical mobile inference latency constraints
4 Method
4.1 Problem Formulation
以前方法的 objective function
m m m 是 model, A C C ACC ACC 是 accuracy, L A T LAT LAT 是 inference latency, T T T 是 target latency
上面的 objective 仅考虑了精度,没有考虑速度
作者 more interested in finding multiple Pareto-optimal solutions in a single architecture search(速度和精度的 trade-off)
设计了如下的 objective function

根据 α \alpha α 和 β \beta β 取值的不同,有如下的 soft 和 hard 版

横坐标是 latency,纵坐标为 objective
soft 版本 − 0.07 -0.07 −0.07 的由来如下:
we empirically observed doubling the latency usually brings about 5% relative accuracy gain
R e w a r d ( M 2 ) = a ⋅ ( 1 + % 5 ) ⋅ ( 2 l / T ) β ≈ R e w a r d ( M 1 ) = a ⋅ ( l / T ) β Reward(M2) = a \cdot (1 + %5 ) \cdot (2l/T )^{\beta}\approx Reward(M1) = a \cdot (l/T )^{\beta} Reward(M2)=a⋅(1+%5)⋅(2l/T)β≈Reward(M1)=a⋅(l/T)β
根据上面公式求出来 β ≈ − 0.07 \beta \approx -0.07 β≈−0.07
4.2 Factorized Hierarchical Search Space

allowing different layer architectures in different blocks
同一个 block 中的 N 个 layer 是一样的,layer 里面的操作如下

搜索的时候 using MobileNetV2 as a reference
每个 layers 数量 {0, +1, -1} based on MobileNetV2
filter size per layer {0.75, 1.0, 1.25} to MobileNetV2
成品结构之一


搜索空间的大小如下:
假设 B B B blocks,and each block has a sub search space of size S S S with average N N N layers per block
搜索空间大小为 S B S^B SB
每个 layer 都不同的话,则为 S B ∗ N S^{B*N} SB∗N
4.3 Search Algorithm

sample-eval-update loop,maximize the expected reward:
reward value R(m) 用的是 objective function
5 Experiments
5.1 Datasets
directly perform our architecture search on the ImageNet training set but with fewer training steps (5 epochs)
区别于 NasNet 的 Cifar10
5.2 Results
1)ImageNet Classification Performance
T = 75 ms,一次搜索,多个 model A1 / A2 / A3
相比 mobileNet v2,引入了 SE 模块,探讨下 SE 模块的影响
2)Model Scaling Performance

这里的 depth multiplier 指的是 channels,可以看出全方位领先 mobilenet v2
作者也可以灵活的通过改变 NAS 时 T 的值来控制模型的大小,上表可以看出,比在大模型上砍通道数效果更猛
3)COCO Object Detection Performance

没什么好评论的,都是菜鸡互啄,哈哈,开玩笑哒,有一定提升
5.3 Ablation Study and Discussion
1)Soft vs. Hard Latency Constraint


hard 版 focus more on faster models to avoid the latency penalty(objective function 也可以看出)
soft 版 tries to search for models across a wider latency range
2)Disentangling Search Space and Reward

解耦探讨下两个创新点的作用
3)Layer Diversity

6 Conclusion(own)
在 mobilenet v2 基础上搜
Pareto-optimal,帕累托最优(来自百度百科)
帕累托最优(Pareto Optimality),也称为帕累托效率(Pareto efficiency),是指资源分配的一种理想状态,假定固有的一群人和可分配的资源,从一种分配状态到另一种状态的变化中,在没有使任何人境况变坏的前提下,使得至少一个人变得更好,这就是帕累托改进或帕累托最优化。
帕累托最优状态就是不可能再有更多的帕累托改进的余地;换句话说,帕累托改进是达到帕累托最优的路径和方法。 帕累托最优是公平与效率的“理想王国”。是由帕累托提出的。
边栏推荐
- 机器学习理论学习:感知机
- Feeling after reading "agile and tidy way: return to origin"
- [introduction to information retrieval] Chapter 6 term weight and vector space model
- ABM thesis translation
- PointNet原理证明与理解
- 程序的内存模型
- 传统目标检测笔记1__ Viola Jones
- 【Cutout】《Improved Regularization of Convolutional Neural Networks with Cutout》
- [multimodal] clip model
- Alpha Beta Pruning in Adversarial Search
猜你喜欢

Interpretation of ernie1.0 and ernie2.0 papers

The difference and understanding between generative model and discriminant model

【信息检索导论】第七章搜索系统中的评分计算

Pointnet understanding (step 4 of pointnet Implementation)

【深度学习系列(八)】:Transoform原理及实战之原理篇

生成模型与判别模型的区别与理解

【Sparse-to-Dense】《Sparse-to-Dense:Depth Prediction from Sparse Depth Samples and a Single Image》
![[paper introduction] r-drop: regulated dropout for neural networks](/img/09/4755e094b789b560c6b10323ebd5c1.png)
[paper introduction] r-drop: regulated dropout for neural networks

@Transitional step pit

【Cutout】《Improved Regularization of Convolutional Neural Networks with Cutout》
随机推荐
Deep learning classification Optimization Practice
Drawing mechanism of view (I)
Record of problems in the construction process of IOD and detectron2
mmdetection训练自己的数据集--CVAT标注文件导出coco格式及相关操作
Feeling after reading "agile and tidy way: return to origin"
Thesis tips
解决latex图片浮动的问题
Conda 创建,复制,分享虚拟环境
一份Slide两张表格带你快速了解目标检测
latex公式正体和斜体
基于pytorch的YOLOv5单张图片检测实现
MMDetection模型微调
Conversion of numerical amount into capital figures in PHP
A slide with two tables will help you quickly understand the target detection
Mmdetection trains its own data set -- export coco format of cvat annotation file and related operations
【信息检索导论】第三章 容错式检索
机器学习理论学习:感知机
【信息检索导论】第一章 布尔检索
[introduction to information retrieval] Chapter 1 Boolean retrieval
【Cascade FPD】《Deep Convolutional Network Cascade for Facial Point Detection》