MushroomRL Benchmark

Reinforcement Learning python library

MushroomRL Benchmark is a benchmarking tool for the Mushroom RL library. The focus of this benchmarking tool is to benchmark the results of deep reinforcement learning algorithms, in particular Deep Actor-Critic. The idea behind MushroomRL Benchmarking is to have a complete platform to run batch comparisons of Deep RL algorithms implemented in MushroomRL under a set of standard benchmark tasks.

With MushroomRL Benchmarking you can:

• Run the benchmarks in a local machine, both sequentially and in parallel fashion

• Run experiments on a SLURM-based cluster.

Download and installation

MushroomRL Benchmark can be downloaded from the GitHub repository. Installation can be done running

cd mushroom-rl-benchmark
pip install -e .[all]

To compile the documentation:

cd mushroom-rl-benchmark/docs
make html

or to compile the pdf version:

cd mushroom-rl-benchmark/docs
make latexpdf

Benchmarks

Benchmarks Results:

• Policy Search Benchmarks
  • Classic Control Environments Benchmarks
• Actor-Critic Benchmarks
  • Classic Control Environments Benchmarks
  • Gym Control Environments Benchmarks
  • Mujoco Environments Benchmarks
  • Bullet Environments Benchmarks
• Value-Based Benchmarks
  • Finite State Environment Benchmark
  • Gym Environments Benchmarks
  • Atari Environment Benchmark

API:

• Core functionality
  • Suite
  • Experiment
  • Logger
  • Visualizer
• Builders
  • EnvironmentBuilder
  • AgentBuilder
  • Policy Search Builders
  • Value Based Builders
  • Actor Critic Builders
• Networks
  • A2CNetwork
  • DDPGCriticNetwork
  • DDPGActorNetwork
  • SACCriticNetwork
  • SACActorNetwork
  • TD3CriticNetwork
  • TD3ActorNetwork
  • TRPONetwork
• Experiment
  • exec_run()
  • compute_metrics()
  • print_metrics()
  • Slurm utilities
• Utils
  • extract_arguments()
  • object_to_primitive()
  • dictionary_to_primitive()
  • build_sweep_list()
  • build_sweep_dict()
  • generate_sweep()
  • generate_sweep_params()