Hi everyone,
Recently I purchased an OAKD-lite camera so I want to have my own model running in it. I have created a dataset and trained it in Roboflow and it serves the purpose amazingly. Then the issue: putting it inside the camera. I have tried multiple tutorials from Luxonis/Roboflow/anything but all are outdated and don't work anymore. I managed to have my blob but the setup is just incredibly frustrating and difficult.
It would be nice to have some basic tutorial about how to use your own model/dataset that is not outdated and uses tensorflow1.x which even google colab has deprecated or yolo3 (altought at this point I would accept anything)
I think my petition is very reasonable considering that its the most interesting feature of the camera. I would like to know is any of you know any resources or similar, otherwise I will be forced to return the camera.
Abrev: I am just using a simple YOLO net to detect 4 classes. Impossible to put it in the camera.