When creating a post, please add:
- Week
I am working on problem beyond the course. If I want to use SSD or YOLO to detect several symbols in one image, How can I train it to do so. How should the input data look like and how should the output layer be like.
I have never personally tried anything like that, but there are some very detailed threads on the forum about YOLO, both how it works and how to train it. Here’s one to start with.
Here’s one that explains how Anchor Boxes work, which are a key part of the algorithm. Here’s another says more about Anchor Boxes and also describes the training dataset.