Representing reputation - turn (21, 20, 3) observations into (21, 20, 4) where the last dimension is 0 wherever there is not an agent, and the normalized reputation value where there is one
@bengreenberg5 @RedTachyon one of the 3 dimensions is already agent. should i just replace the 1 with the reputation score from the ledger or should we really have a whole nother dimension?
Representing reputation - turn
(21, 20, 3)
observations into(21, 20, 4)
where the last dimension is 0 wherever there is not an agent, and the normalized reputation value where there is onethe normalized value will come from https://github.com/RedTachyon/cpr_reputation/issues/33