I have a yolo Darknet file generated from a custom, fairly small, dataset. I used LabelImg & [login to view URL] (they offer the ability to export in nearly any format, so this can easily change if need be) to generate it.
There are TWO classes in this dataset. In my experimentation, 6000 batches in Google's colab on a tesla p100 took approximately 3 hours. I however have no idea what I'm doing, and blindly following guides. Please help.
I'd love to send you the zip file or download code in order to pull my dataset, and train an object detection model on it, and have the ability to run on TFJS.
This project may have future offers as well. Please feel free to ask any questions about anything I may have been unclear on.
Upon completion of this project, I need to be able to spin up a TFJS instance, open the web cam, and show bounding boxes around the 2 classes of objects provided.