You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At present, only the local_only neuron receivers can handle 2D input (and only the local_only_pool_dense connector can cope with nD data). Ideally, normal populations would be able to do this to.
For this to work, the population table needs to be updated to cope with the 2D (or nD) structure of the key, which will have fields for each dimension (e.g. for a 2D population split into 2x3 squares, the key will have a 1-bit field for y and a 2-bit field for x, which is why it doesn't just make a linear key). If nD is considered, the information will have to be more dynamic in size and different for each incoming population, probably akin to the address list.
The text was updated successfully, but these errors were encountered:
At present, only the local_only neuron receivers can handle 2D input (and only the local_only_pool_dense connector can cope with nD data). Ideally, normal populations would be able to do this to.
For this to work, the population table needs to be updated to cope with the 2D (or nD) structure of the key, which will have fields for each dimension (e.g. for a 2D population split into 2x3 squares, the key will have a 1-bit field for y and a 2-bit field for x, which is why it doesn't just make a linear key). If nD is considered, the information will have to be more dynamic in size and different for each incoming population, probably akin to the address list.
The text was updated successfully, but these errors were encountered: