Why can’t I see FOMO option at neural network architecture? There is mobilenets only
Project ID:
Context/Use case:
Hi @formashin,
Go to your Dashboard on your project. Under Project info, change Labeling method to Bounding boxes (object detection). From there, you will probably need to recreate your impulse. For the Learning block, you should see the option to use an object detection model. Then, in the learning block page, you should be able to select an object detection model, such as FOMO or MobileNet-SSD.
Thanks for reply.
I didn’t pay attention than FOMO is for object detection tasks only.
Why isn’t FOMO possible use for image classification
Hi @formashin,
FOMO is a model architecture that was built specifically for constrained object detection. It divides the image into a grid and attempts to classify each grid cell as belonging to an object or not. I recommend checking out my talk here if you would like to learn more about how FOMO works: https://www.youtube.com/watch?v=VzJZM5p24Tc
If you want to do image classification, I recommend either a 3-layer CNN (for simple tasks) or MobileNetv2 (for more complex tasks).