Skip to content

Latest commit

 

History

History
executable file
·
24 lines (15 loc) · 1.15 KB

README.md

File metadata and controls

executable file
·
24 lines (15 loc) · 1.15 KB

YOLOv8 with custom dataset and configuration

YOLOv8 is a deep learning model developed by Ultralytics. More about YOLOv8 and Ultralytics can be found here. YOLOv8 can be installed with:

pip install ultralytics

Dataset

All about preparing the YOLOv8 dataset and what it should look like can be found here.

N.B. Proper preparation of the dataset is one of the most important parts for successful model learning! You should not skip it out and always check if your dataset meets the YOLOv8 conditions!

Custom configuration

There is a defualt configuration (defualt.yaml) in this repository that you should use as an example. Each argument in default.yaml is described here.

I suggest that you create a custom configuration file that suits your needs.

Training, validation, prediction, ...

All this can be set in the configuration file in the mode argument. After training, all used arguments, diagrams, results and weights are saved in the ./project/name directory.

TODO

  • add tensorboard or clearML