Crate gmgn 

gmgn — Rust Reinforcement Learning Gymnasium

A pure-Rust library for reinforcement learning environments, inspired by Gymnasium. Designed for zero-cost abstraction and idiomatic Rust ergonomics.

Quick Start (static, generic)

use gmgn::prelude::*;
use gmgn::envs::classic_control::{CartPoleEnv, CartPoleConfig};

let mut env = CartPoleEnv::new(CartPoleConfig::default()).unwrap();
let reset = env.reset(Some(42)).unwrap();

for _ in 0..200 {
    let action = env.action_space().sample(&mut gmgn::rng::create_rng(None));
    let step = env.step(&action).unwrap();
    if step.terminated {
        break;
    }
}

Quick Start (dynamic, registry)

use gmgn::registry::{self, DynValue};

registry::register_builtins();
let mut env = registry::make("CartPole-v1").unwrap();
env.reset_dyn(Some(42)).unwrap();
let step = env.step_dyn(&DynValue::Discrete(1)).unwrap();

Modules

env

Core environment trait and associated types.

env_checker

Environment validation utilities.

envs

Built-in environment implementations.

error

Error types for the gmgn reinforcement learning library.

prelude

Convenience re-exports for common gmgn types.

registry

Global environment registry and factory functions.

render

2D rendering system for environment visualization.

rng

Random number generation for reproducible environments.

space

Observation and action space definitions.

vector

Vectorized environments for batched parallel execution.

wrappers

Environment wrappers for modular behavior transformation.