Linux founder Linus Torvalds famously said:. Talk is cheap Show me the code (redundant to talk about enough, grading up!).
The code is read from the entry to improve the only way. Especially for deep learning, many hides the implementation framework underlying neural network, using only the upper transfer package, it is difficult to know the inner workings of clear, is not conducive to further optimization and innovation.
YOLOv3 end is a real-time target detection method based on the depth study to known speed.
YOLOv3 implementation Darknet is the use of lightweight open source framework for deep learning C language development, less dependent, portability, and can serve as a good code to read the case, let us delve into its implementation principle.
Taking into account this aspect needs, we introduced the course "YOLOv3 target detection: Principles and parse the source code."
Course Link: https://edu.51cto.com/course/18384.html
This course will resolve YOLOv3 the principle and source code, the specific content includes:
-
YOLO target detection principle
-
Neural networks and Darknet C language, and solving the gradient back propagation of error is calculated in particular
-
The code reading tools and methods
-
Depth study calculated weapon: BLAS and GEMM
-
GPU's CUDA programming methods and applications in Darknet
- Program flow YOLOv3 and resolve the source of each layer
Source files Darknet after this course will provide comments.
In addition to this course, "YOLOv3 target detection: Principles and source code analysis" outside, I launched a series of courses on YOLOv3 target detection, including:
"YOLOv3 target detection combat: training their own data set"
"YOLOv3 target detection combat: traffic sign recognition."
"YOLOv3 target detection: Principles and source code analysis"
"YOLOv3 target detection: network model Improvement"
First proposed course of study "YOLOv3 target detection combat: training their own data sets," or course "YOLOv3 target detection combat: traffic sign recognition", then this course for the future use of YOLOv3 understanding.