Does fomo algorithm supports gpu for object detection if so can you explain how to use it. Also how to trained the model on gpu and how to generate the code for it.
Hello @thisisKaustubh,
GPU training is reserved for our Enterprise users.
Here is a tutorial on how to train a FOMO model: Detect objects with centroids - Edge Impulse Documentation
FOMO should be able to run on GPU but it’s already super fast on CPU so I don’t think that should be necessary. Happy to hear about your use-case.
Do not hesitate if you have any other questions.
Best,
Louis
Thanks for the reply!
Also I would like to know if we build our project in deployment section as a C++ library then what is the default model version(uint8 or float32) with for eg FOMO MNV2 0.1?
Basically I just want a float32 version as I get faster inference time(35ms) there, so need to know how do we export the float32 version only.
Thanks for your time in advance!
Hello @thisisKaustubh,
When you download your model, you can select which model you want.
Best,
Louis
Thanks for the help and your time!
1 Like