[paper]
[implementation]
Ape-X variations of DQN, DDPG, and QMIX (APEX_DQN, APEX_DDPG, APEX_QMIX) use a single GPU learner and many CPU workers for experience collection. Experience collection can scale to hundreds of CPU workers due to the distributed prioritization of experience prior to storage in replay buffers.
Ape-X architecture
Tuned examples: PongNoFrameskip-v4, Pendulum-v0, MountainCarContinuous-v0, {BeamRider,Breakout,Qbert,SpaceInvaders}NoFrameskip-v4.
Atari results @10M steps: more details
| Atari env | RLlib Ape-X 8-workers | Mnih et al Async DQN 16-workers |
|---|---|---|
| BeamRider | 6134 | ~6000 |
| Breakout | 123 | ~50 |
| Qbert | 15302 | ~1200 |
| SpaceInvaders | 686 | ~600 |
Scalability:
| Atari env | RLlib Ape-X 8-workers @1 hour | Mnih et al Async DQN 16-workers @1 hour |
|---|---|---|
| BeamRider | 4873 | ~1000 |
| Breakout | 77 | ~10 |
| Qbert | 4083 | ~500 |
| SpaceInvaders | 646 | ~300 |
Ape-X using 32 workers in RLlib vs vanilla DQN (orange) and A3C (blue) on PongNoFrameskip-v4.
Ape-X specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/dqn/apex.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper]
[implementation]
In IMPALA, a central learner runs SGD in a tight loop while asynchronously pulling sample batches from many actor processes. RLlib's IMPALA implementation uses DeepMind's reference V-trace code. Note that we do not provide a deep residual network out of the box, but one can be plugged in as a custom model. Multiple learner GPUs and experience replay are also supported.
IMPALA architecture
Tuned examples: PongNoFrameskip-v4, vectorized configuration, multi-gpu configuration, {BeamRider,Breakout,Qbert,SpaceInvaders}NoFrameskip-v4
Atari results @10M steps: more details
| Atari env | RLlib IMPALA 32-workers | Mnih et al A3C 16-workers |
|---|---|---|
| BeamRider | 2071 | ~3000 |
| Breakout | 385 | ~150 |
| Qbert | 4068 | ~1000 |
| SpaceInvaders | 719 | ~600 |
Scalability:
| Atari env | RLlib IMPALA 32-workers @1 hour | Mnih et al A3C 16-workers @1 hour |
|---|---|---|
| BeamRider | 3181 | ~1000 |
| Breakout | 538 | ~10 |
| Qbert | 10850 | ~500 |
| SpaceInvaders | 843 | ~300 |
Multi-GPU IMPALA scales up to solve PongNoFrameskip-v4 in ~3 minutes using a pair of V100 GPUs and 128 CPU workers. The maximum training throughput reached is ~30k transitions per second (~120k environment frames per second).
IMPALA-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/impala/impala.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper]
[implementation]
We include an asynchronous variant of Proximal Policy Optimization (PPO) based on the IMPALA architecture. This is similar to IMPALA but using a surrogate policy loss with clipping. Compared to synchronous PPO, APPO is more efficient in wall-clock time due to its use of asynchronous sampling. Using a clipped loss also allows for multiple SGD passes, and therefore the potential for better sample efficiency compared to IMPALA. V-trace can also be enabled to correct for off-policy samples.
Tip
APPO is not always more efficient; it is often better to use standard PPO or IMPALA.
APPO architecture (same as IMPALA)
Tuned examples: PongNoFrameskip-v4
APPO-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/ppo/appo.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper]
[implementation]
Unlike APPO or PPO, with DD-PPO policy improvement is no longer done centralized in the trainer process. Instead, gradients are computed remotely on each rollout worker and all-reduced at each mini-batch using torch distributed. This allows each worker's GPU to be used both for sampling and for training.
Tip
DD-PPO is best for envs that require GPUs to function, or if you need to scale out SGD to multiple nodes. If you don't meet these requirements, standard PPO will be more efficient.
DD-PPO architecture (both sampling and learning are done on worker GPUs)
Tuned examples: CartPole-v0, BreakoutNoFrameskip-v4
DDPPO-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/ppo/ddppo.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation]
RLlib implements A2C and A3C using SyncSamplesOptimizer and AsyncGradientsOptimizer respectively for policy optimization. These algorithms scale to up to 16-32 worker processes depending on the environment.
A2C also supports microbatching (i.e., gradient accumulation), which can be enabled by setting the microbatch_size config. Microbatching allows for training with a train_batch_size much larger than GPU memory. See also the microbatch optimizer implementation.
A2C architecture
Tuned examples: PongDeterministic-v4, PyTorch version, {BeamRider,Breakout,Qbert,SpaceInvaders}NoFrameskip-v4
Tip
Consider using IMPALA for faster training with similar timestep efficiency.
Atari results @10M steps: more details
| Atari env | RLlib A2C 5-workers | Mnih et al A3C 16-workers |
|---|---|---|
| BeamRider | 1401 | ~3000 |
| Breakout | 374 | ~150 |
| Qbert | 3620 | ~1000 |
| SpaceInvaders | 692 | ~600 |
A3C-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/a3c/a3c.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation]
DDPG is implemented similarly to DQN (below). The algorithm can be scaled by increasing the number of workers, switching to AsyncGradientsOptimizer, or using Ape-X. The improvements from TD3 are available as TD3.
DDPG architecture (same as DQN)
Tuned examples: Pendulum-v0, MountainCarContinuous-v0, HalfCheetah-v2, TD3 Pendulum-v0, TD3 InvertedPendulum-v2, TD3 Mujoco suite (Ant-v2, HalfCheetah-v2, Hopper-v2, Walker2d-v2).
DDPG-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/ddpg/ddpg.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation]
RLlib DQN is implemented using the SyncReplayOptimizer. The algorithm can be scaled by increasing the number of workers, using the AsyncGradientsOptimizer for async DQN, or using Ape-X. Memory usage is reduced by compressing samples in the replay buffer with LZ4. All of the DQN improvements evaluated in Rainbow are available, though not all are enabled by default. See also how to use parametric-actions in DQN.
DQN architecture
Tuned examples: PongDeterministic-v4, Rainbow configuration, {BeamRider,Breakout,Qbert,SpaceInvaders}NoFrameskip-v4, with Dueling and Double-Q, with Distributional DQN.
Tip
Consider using Ape-X for faster training with similar timestep efficiency.
Hint
For a complete rainbow setup,
make the following changes to the default DQN config:
"n_step": [between 1 and 10],
"noisy": True,
"num_atoms": [more than 1],
"v_min": -10.0,
"v_max": 10.0
(set v_min and v_max according to your expected range of returns).
Atari results @10M steps: more details
| Atari env | RLlib DQN | RLlib Dueling DDQN | RLlib Dist. DQN | Hessel et al. DQN |
|---|---|---|---|---|
| BeamRider | 2869 | 1910 | 4447 | ~2000 |
| Breakout | 287 | 312 | 410 | ~150 |
| Qbert | 3921 | 7968 | 15780 | ~4000 |
| SpaceInvaders | 650 | 1001 | 1025 | ~500 |
DQN-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/dqn/dqn.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation] We include a vanilla policy gradients implementation as an example algorithm.
Policy gradients architecture (same as A2C)
Tuned examples: CartPole-v0
PG-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/pg/pg.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation]
PPO's clipped objective supports multiple SGD passes over the same batch of experiences. RLlib's multi-GPU optimizer pins that data in GPU memory to avoid unnecessary transfers from host memory, substantially improving performance over a naive implementation. PPO scales out using multiple workers for experience collection, and also to multiple GPUs for SGD.
Tip
If you need to scale out with GPUs on multiple nodes, consider using decentralized PPO.
PPO architecture
Tuned examples: Humanoid-v1, Hopper-v1, Pendulum-v0, PongDeterministic-v4, Walker2d-v1, HalfCheetah-v2, {BeamRider,Breakout,Qbert,SpaceInvaders}NoFrameskip-v4
Atari results: more details
| Atari env | RLlib PPO @10M | RLlib PPO @25M | Baselines PPO @10M |
|---|---|---|---|
| BeamRider | 2807 | 4480 | ~1800 |
| Breakout | 104 | 201 | ~250 |
| Qbert | 11085 | 14247 | ~14000 |
| SpaceInvaders | 671 | 944 | ~800 |
Scalability: more details
| MuJoCo env | RLlib PPO 16-workers @ 1h | Fan et al PPO 16-workers @ 1h |
|---|---|---|
| HalfCheetah | 9664 | ~7700 |
RLlib's multi-GPU PPO scales to multiple GPUs and hundreds of CPUs on solving the Humanoid-v1 task. Here we compare against a reference MPI-based implementation.
PPO-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/ppo/ppo.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
SAC architecture (same as DQN)
RLlib's soft-actor critic implementation is ported from the official SAC repo to better integrate with RLlib APIs. Note that SAC has two fields to configure for custom models: policy_model and Q_model, and currently has no support for non-continuous action distributions.
Tuned examples: Pendulum-v0, HalfCheetah-v3
MuJoCo results @3M steps: more details
| MuJoCo env | RLlib SAC | Haarnoja et al SAC |
|---|---|---|
| HalfCheetah | 13000 | ~15000 |
SAC-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/sac/sac.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation]
ARS is a random search method for training linear policies for continuous control problems. Code here is adapted from https://github.com/modestyachts/ARS to integrate with RLlib APIs.
Tuned examples: CartPole-v0, Swimmer-v2
ARS-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/ars/ars.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation]
Code here is adapted from https://github.com/openai/evolution-strategies-starter to execute in the distributed setting with Ray.
Tuned examples: Humanoid-v1
Scalability:
RLlib's ES implementation scales further and is faster than a reference Redis implementation on solving the Humanoid-v1 task.
ES-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/es/es.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation] Q-Mix is a specialized multi-agent algorithm. Code here is adapted from https://github.com/oxwhirl/pymarl_alpha to integrate with RLlib multi-agent APIs. To use Q-Mix, you must specify an agent grouping in the environment (see the two-step game example). Currently, all agents in the group must be homogeneous. The algorithm can be scaled by increasing the number of workers or using Ape-X.
Tuned examples: Two-step game
QMIX-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/qmix/qmix.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation] MADDPG is a specialized multi-agent algorithm. Code here is adapted from https://github.com/openai/maddpg to integrate with RLlib multi-agent APIs. Please check justinkterry/maddpg-rllib for examples and more information.
MADDPG-specific configs (see also common configs):
Tuned examples: Multi-Agent Particle Environment, Two-step game
.. literalinclude:: ../../rllib/contrib/maddpg/maddpg.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation] MARWIL is a hybrid imitation learning and policy gradient algorithm suitable for training on batched historical data. When the beta hyperparameter is set to zero, the MARWIL objective reduces to vanilla imitation learning. MARWIL requires the offline datasets API to be used.
Tuned examples: CartPole-v0
MARWIL-specific configs (see also common configs):
.. literalinclude:: ../../rllib/agents/marwil/marwil.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__
[paper] [implementation] AlphaZero is an RL agent originally designed for two-player games. This version adapts it to handle single player games. The code can be used with the SyncSamplesOptimizer as well as with a modified version of the SyncReplayOptimizer, and it scales to any number of workers. It also implements the ranked rewards (R2) strategy to enable self-play even in the one-player setting. The code is mainly purposed to be used for combinatorial optimization.
Tuned examples: CartPole-v0
AlphaZero-specific configs (see also common configs):
.. literalinclude:: ../../rllib/contrib/alpha_zero/core/alpha_zero_trainer.py :language: python :start-after: __sphinx_doc_begin__ :end-before: __sphinx_doc_end__