#reinforcement-learning #game-theory #model #player #env-var #data-model #theory

amfiteatr_rl

Extensions providing reinforcement learning implementations for amfiteatr

7 releases (4 breaking)

0.5.0 Apr 9, 2024
0.4.1 Apr 2, 2024
0.3.0 Mar 11, 2024
0.2.1 Feb 7, 2024
0.1.0 Jan 15, 2024

#391 in Simulation


Used in 3 crates

Custom license

300KB
5K SLoC

AmfiRL

Framework library for reinforcement learning using data model from amfi crate. Crate contains traits and generic implementation.

Torch dependency

This crate depends on tch which requires that you have Torch installed. Use guide provided by that crate.

For me the procedure looks like this:

  1. Unpacking compiled torchlib for C++/Java; You will get structure like this:
/path/to/libtorch
                | -- bin
                | -- include
                | -- lib
                | ...
  1. Set your environment variables:
export LIBTORCH=/path/to/torch
export LD_LIBRARY_PATH=/path/to/libtorch/lib

Examples

Examples are presented in separate crate - amfiteatr_examples

Licence: MIT


lib.rs:

AmfiRL

Framework library for reinforcement learning using data model from amfiteatr. Crate contains traits and generic implementation.

Examples

Examples are presented in separate crate - amfiteatr_examples

Licence: MIT

Dependencies

~13MB
~262K SLoC