nengo / nengo-loihi

Run Nengo models on Intel's Loihi chip
https://www.nengo.ai/nengo-loihi/
Other
35 stars 12 forks source link

Avoid full weights for scalar transforms #232

Closed hunse closed 5 years ago

hunse commented 5 years ago

Right now, we can't do something like nengo.Connection(u, a.neurons) if the number of neurons is large, even though this connection doesn't have many weights. This is because we make a full identity transform in expand_to_2d. We should get rid of this function, since it's easy to do on Loihi.

drasmuss commented 5 years ago

Done in #226