openai / mlsh

Code for the paper "Meta-Learning Shared Hierarchies"

Status: Archive (code is provided as-is, no updates expected)

Meta-Learning Shared Hierarchies

Code for Meta-Learning Shared Hierarchies.

Installation

Add to your .bash_profile (replace ... with path to directory):
export PYTHONPATH=$PYTHONPATH:/.../mlsh/gym;
export PYTHONPATH=$PYTHONPATH:/.../mlsh/rl-algs;

Install MovementBandits environments:
cd test_envs
pip install -e .

Running Experiments

python main.py --task AntBandits-v1 --num_subs 2 --macro_duration 1000 --num_rollouts 2000 --warmup_time 20 --train_time 30 --replay False AntAgent

Once you've trained your agent, view it by running:

python main.py [...] --replay True --continue_iter [your iteration] AntAgent

The MLSH script works on any Gym environment that implements the randomizeCorrect() function. See the envs/ folder for examples of such environments.

To run on multiple cores:

mpirun -np 12 python main.py ...