All activity
Mo Assafleft a comment
As a developer building YOLO models all the time, I found it extremely stupid to keep labelling images manually. There are many vision datasets and tools out there to develop vision models. And yet it remains a pain to find a suitable dataset or to create a new one from scratch. It needs to look like your data or match your use case, and be of high quality. It's time to automate all this. Help...
FissionBox, Early Beta AccessBuild vision/YOLO models fast — without labelling data
Simply upload footage, give the vision reactor a goal (e.g., "track cat faces"), click "fission", that's it! Now you download a labelled YOLO dataset and/or train a model.
Built for devs, data scientists, and startups moving fast.
Bye labelling!
FissionBox, Early Beta AccessBuild vision/YOLO models fast — without labelling data