Hacker News Reader

> Ok how the hell do we train this thing? Stochastic gradient descent with back-propagation won't work here (or if it does I have no idea how to implement it).

What's wrong with gradient descent?

https://snntorch.readthedocs.io/en/latest/

View on HN

Thanks for sharing. I thought the discontinuous nature of the SNN made it non-differentiable and therefore unsuitable for SGD and backprop.

Lol in differentiable programming they usually hard code an identity for the problematic parts (e.g. if statements)

Gradient descent needs a differentiable system, the author's clearly not.

Reply on news.ycombinator.com