MVIG-SJTU / AlphAction

Spatio-Temporal Action Localization System
407 stars 74 forks source link

作者您好,拜读了您的作品,有几个问题想请教您。希望可以得到您的解答。 #44

Closed wen-flow closed 3 years ago

wen-flow commented 3 years ago

yolo 检测器和 Tracker 的模型文件是如何训练出来的啊, 我该如何训练自己的检测器和追踪模型呢?

yelantf commented 3 years ago

The model used in the demo program is simply imported from other programs, you could refer to the original repository for more information. For the training and inference, we use maskrcnn-benchmark to train the human detectors, you can refer to the final step in DATA.md and maskrcnn-benchmark for more details. Besides, you may also interested in this issue https://github.com/MVIG-SJTU/AlphAction/issues/40 too.