The notebook
You probably want to look here.
Example
Things you need
Download the weights:
wget http://pjreddie.com/media/files/yolo.weightsQuestions
- Why does this particular set of layers work so well together.
- How does the loss function work
most questions are at the top of sections in the notebook
Credits
This project was heavily inspired by this repo and this article and obviosly is based on the yolo network.
Random
miliseconds per evaluation: ~350
you can see the model images in the doc directory.
Roadmap
- Implement layers of network (Q1)
- Pipeline for image processing
- Fix weight reader
- Once we are here we should be able to just load any config file and have it work
- Make config reader "standard"
- Understand loss function (Q2)
- Implementing loss function and training
- Improve upon network
- Documentation of network