cyclesgym

Cyclesgym

This repository contains an OpenAI gym interface to the Cycles crop growth simulator.

For more information about cyclesgym, see our user manual.

Installation

We recommend Python 3.8+ installation using Anaconda.

First, create and activate a virtual environment using Anaconda:

conda create -yn cyclesgym python=3.8
conda activate cyclesgym

Then, clone the repo and change working directory

git clone https://github.com/kora-labs/cyclesgym.git
cd cyclesgym

Subsequently, install the library according to your needs. To install, run:

pip install -e .

If you further want to use some basic libraries to train reinforcement learning agents of the cyclesgym environments use:

pip install -e .SOLVERS

Or, if you are using zsh:

pip install -e .\[SOLVERS\]

Contextual MDP experiments

We provide a set of scripts to run contextual MDP experiments on the cyclesgym environments with crop planning tasks.

To train the policy and evaluate on CMDP, run the following command:

cd experiments/crop_planning
sbatch s-crop-train-transfer.sh

For multi-task training, run:

sbatch s-crop-train-transfer-multi.sh

The results will be saved in the results directory.