The latest dataset

New dataset

Second Release of Models!

valorant-v8.pt and v9 are almost equal.

Why is it almost equal? Because Google Colab wasn't able to process due to a lot of images.
So I changed batch-size 80 to 50. It turned out to be almost the same.

If anyone knows how to process 4126 Train Images and 2675 Val Images in Google Colab Pro with batch-size 80, Please let me know.

Full Changelog: model-v1.0...model-v2.0

The first release of Valorant models trained in Yolov5!

This is a pack that is included from valorant-n-1.pt to valorant-n-3.pt and from valorant-v1.pt to valorant-v8.pt. Plus, It's included krunker.pt :D

Please give me a review. Which model is the best precise for example. ( I think the best is v8 trained in yolov5s)

Full Changelog: https://github.com/Leaf48/YOLOv5-Models-For-Valorant/commits/model-v1.0

About 2000 images were added to the original dataset.

Thanks for contributing many images, Merts!

Full Changelog: dataset-v1.0...dataset-v2.0

Release Dataset

This dataset was used in valorant-n and valorant-v series.

I labeled "enemy" from the knees to the top of the head.