Closed dev2y closed 1 year ago
Not yet.
@jeshraghian Thank you. But it is possible to do it. Right?
Definitely! The best way to get started is to use a pre existing PyTorch implementation of R-CNN. Then between each Conv2d/Linear layer, include a Leaky
neuron from snntorch.
You can pass the spikes from the Leaky layer to the pytorch layers.
It might struggle to learn anything to start with, so you may need to play around with optimizers, surrogate gradients, and lowering firing thresholds. It'll be a little tedious, but the discord channel is the best place to get help.
Thank you. I will give it a try.
Hi is there any training example for R-CNN? Thanks.