Titanfall 2 - Enemy detection using computer vision (V1)

Описание к видео Titanfall 2 - Enemy detection using computer vision (V1)

This model was trained on a custom data set gathered from gameplay of Titanfall 2, it was trained on a 2060 using yoloV8, the training set had roughly 2700 images and took 12 hours to train to this level, there are still a lot of things I want to improve on such as individual enemy classification and I also want to improve the ai confidence % so random objects stop being detected.

Комментарии

Информация по комментариям в разработке