What the hell to write about the new tech column? This problem has troubled our engineers for a long time. Regarding deep learning, there are too many materials and documents that can be consulted on the Internet. As long as you are willing to learn, you can learn everything from entry to mastery. Until one day, a friend in our AI technology exchange group asked if he could share some practical cases so that those who are interested and have time can learn and communicate in practice.
This is indeed a good direction. What is achieved on paper is always shallow, and I know that this matter must be done. Theoretical knowledge is too much and too complicated, but it is difficult to internalize and refine it if you can’t see it in practice. Besides, there is no so-called big cow in the world. This column will start with specific cases through some practical projects, and discuss with you in practice.
Column Outline
In the early stage of the column, we will present three practical projects in computer vision. A brief introduction to this single project is given below.
Understanding Neural Networks Using Visualization
It will introduce how to use the visualization method to intuitively understand how the convolutional neural network works through GradCam:
Thinking: How does the convolutional neural network complete the task?
Directory detection based on lightweight network
Introduce the basic principle of target SSD
Introduction to Lightweight Models
Implement a generic object detector
Improvement scheme and experiment of face detection based on SSD
GAN Dafa and its interesting applications
Fundamentals of GANs
CycleGAN realizes the mutual conversion of pictures in two fields
Image transfer between multiple fields of StarGAN
Font Style Migration
Of course, we will update some other project practices in the future. Of course, we welcome your interested directions.
Column features
Valuable experience plus thought collision. In the process of practice, you will definitely encounter some problems and thoughts. These are what the author most wants to share with you through the column. This column will share with you the pitfalls encountered in the project practice and how to climb the pits and other valuable experience. At the same time, we will also solicit suggestions and thoughts after each project. For example, whether there is a better way to achieve model optimization, or other methods to achieve the same effect, or whether the same method can be used to do more interesting things, etc. Perhaps, only in such exchanges and collisions can more progress be brought about.
practice preparation
If you have friends who are interested in practicing together, you need to make the following preparations:
1、熟悉python编程；
2、熟悉深度学习框架pytorch，这些项目我们将主要基于pytorch进行实验。
如果有相关的建议和想法，欢迎添加我们的技术交流群交流哦～添加技术助理微信入群：geetest1024