Module ilpyt.nets
Wrappers for the torch.nn.Module
class, and can be used to parameterize policy
functions, value functions, cost functions, etc. for use in reinforcement
learning and imitation learning algorithms.
To create a custom network, see BaseNetwork
for more details.
Expand source code
"""
Wrappers for the `torch.nn.Module` class, and can be used to parameterize policy
functions, value functions, cost functions, etc. for use in reinforcement
learning and imitation learning algorithms.
To create a custom network, see `BaseNetwork` for more details.
"""
Sub-modules
ilpyt.nets.base_net
-
BaseNetwork
is the abstract class for a network. Networks parameterize important functions during learning - most often, the agent policy … ilpyt.nets.net1d
-
These 1D networks are suited for 1D inputs.
ilpyt.nets.net2d
-
These 2D networks are suited for 3D inputs (h,w,c).