- Added a built-in objective to assist in applying regularization during training.
- Added keep_history config option, which can be set to
Falseon Probes if only the data from the most recent simulation step is desired (as opposed to the default behaviour of keeping the data from all steps).
sim.losswill now apply
nengo_dl.objectives.mseto all probes in
dataif no explicit
objectiveis given (mirroring the default behaviour in
- The Spaun benchmark network will now be installed through pip rather than manually cloning and importing the repo.
- Fixed objective argument parsing if objective is a callable class or method.
- Fixed bug in
sim.train1-step synapse warning when explicitly specifying
n_steps(rather than passing in
"mse"as the objective in
sim.lossis no longer supported. Use the function