Example Environment
This project provides a simple example of how to use the Env interface to generate inputs for a network.
The deep_fsa
and sir2
projects also have Env implementations, but they are more complex, so this example may provide a more generic starting point for building your own custom Env.
The Env code is entirely in env.go
, and sim.go
is just the ra25
simulation code adapted to use the env.
The env just activates a point in a 2D input matrix and has one-hot output representations for the X and Y coordinates of that point.
See also https://github.com/CompCogNeuro/sims for other models with Env implementations, including ch9/sg
which uses the general-purpose emergent/esg
stochastic generator to generate sentence inputs.