DIY Custom Object Detection Model via Transfer Learning (Tensorflow Lite Edge TPU)

Описание к видео DIY Custom Object Detection Model via Transfer Learning (Tensorflow Lite Edge TPU)

A tutorial on how to create your own Tensorflow Lite custom object detection model via transfer learning for Edge TPU devices like the Google Coral and Asus Tinker Edge T.

We'll use this website as a guide: https://coral.ai/docs/edgetpu/retrain...

Chapters:
0:16 Install Docker
1:19 Clone Google Coral Tutorials repo
1:53 Start the Google Coral Tutorials Docker Container
2:11 Download Google's Training Data
3:16 Configure Google's Training Data
4:50 Download & Configure YOUR OWN Training Data
9:26 Copy YOUR OWN Training Data Into The Docker Container
12:05 Start Training YOUR OWN Data
12:45 Monitor Training Progress
13:14 How Much Training Should You Do?
13:40 Compile The Model For The Edge TPU
16:02 Transfer The Model To The Tinker Edge T
17:05 Run The Model
17:27 Outro

Github files: https://github.com/edgecate/ANAL/tree...

(Sorry for the poor audio - OBS on Ubuntu didn't play nice)

Комментарии

Информация по комментариям в разработке