Skip to content

Commit

Permalink
chore: optimize doc
Browse files Browse the repository at this point in the history
  • Loading branch information
HarleysZhang committed Feb 20, 2023
1 parent c4c2c45 commit 5f24264
Show file tree
Hide file tree
Showing 2 changed files with 6 additions and 7 deletions.
6 changes: 1 addition & 5 deletions 5-deep_learning/经典backbone详解/ResNet网络详解.md
Original file line number Diff line number Diff line change
Expand Up @@ -55,11 +55,7 @@ Residual Block(残差块)之间的衔接,在原论文中,$F(x)+x$ 是经

## 二,ResNet18 模型结构分析

残差网络中,将堆叠的几层卷积 `layer` 称为残差块(`Residual Block`),多个相似的残差块串联构成 ResNet。

ResNet18 和 ResNet34 Backbone用的都是两层的残差学习单元(`basic block`),更深层的ResNet则使用的是三层的残差学习单元(`bottle block`)。

ResNet18 模型的卷积层数是 $18$ 层,对应的残差学习 `Blocks` 的单元数量分别为 2、2、2 和 2。
残差网络中,将堆叠的几层卷积 `layer` 称为残差块(`Residual Block`),多个相似的残差块串联构成 ResNet。ResNet18 和 ResNet34 Backbone用的都是两层的残差学习单元(`basic block`),更深层的ResNet则使用的是三层的残差学习单元(`bottle block`)。

ResNet18 其结构如下图所示。

Expand Down
7 changes: 5 additions & 2 deletions cv算法工程师成长路线.md
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@
- [6.5,语义分割文章](#65语义分割文章)
- [6.6,3D 视觉技术文章](#663d-视觉技术文章)
- [6.7,深度学习的评价指标文章](#67深度学习的评价指标文章)
- [七,模型压缩与量化](#七模型压缩与量化)
- [七,模型压缩](#七模型压缩)
- [7.1,轻量级网络设计](#71轻量级网络设计)
- [7.2,模型压缩文章](#72模型压缩文章)
- [7.3,神经网络量化文章](#73神经网络量化文章)
Expand Down Expand Up @@ -221,7 +221,10 @@
4. [PLASTER:一个与深度学习性能有关的框架](https://zhuanlan.zhihu.com/p/38315051)
5. [The Correct Way to Measure Inference Time of Deep Neural Networks](https://towardsdatascience.com/the-correct-way-to-measure-inference-time-of-deep-neural-networks-304a54e5187f)

## 七,模型压缩与量化
## 七,模型压缩

模型压缩与模型推理系统的知识,推荐看微软的[《人工智能系统》](https://microsoft.github.io/AI-System/)教程资料,写的很全且容易理解,个人觉得是个非常不错的教程。

### 7.1,轻量级网络设计

1. [轻量卷积神经网络的设计](https://zhuanlan.zhihu.com/p/64400678 "轻量卷积神经网络的设计")
Expand Down

0 comments on commit 5f24264

Please sign in to comment.