# Proximal Policy Gradient (PPO)
## Overview
PPO is one of the most popular DRL algorithms. It runs reasonably fast by leveraging vector (parallel) environments and naturally works well with different action spaces, therefore supporting a variety of games. It also has good sample efficiency compared to algorithms such as DQN.
Original paper:
* [Proximal Policy Optimization Algorithms](https://arxiv.org/abs/1707.06347)
Reference resources:
* [Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO](https://arxiv.org/abs/2005.12729)
* [What Matters In On-Policy Reinforcement Learning? A Large-Scale Empirical Study](https://arxiv.org/abs/2006.05990)
* ⭐ [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/)
All our PPO implementations below are augmented with the same code-level optimizations presented in `openai/baselines`'s [PPO](https://github.com/openai/baselines/tree/master/baselines/ppo2). To achieve this, see how we matched the implementation details in our blog post [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/).
## Implemented Variants
| Variants Implemented | Description |
| ----------- | ----------- |
| :material-github: [`ppo.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppopy) | For classic control tasks like `CartPole-v1`. |
| :material-github: [`ppo_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_ataripy) | For Atari games. It uses convolutional layers and common atari-based pre-processing techniques. |
| :material-github: [`ppo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_continuous_actionpy) | For continuous action space. Also implemented Mujoco-specific code-level optimizations. |
| :material-github: [`ppo_atari_lstm.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_lstmpy) | For Atari games using LSTM without stacked frames. |
| :material-github: [`ppo_atari_envpool.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpoolpy) | Uses the blazing fast Envpool Atari vectorized environment. |
| :material-github: [`ppo_atari_envpool_xla_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy) | Uses the blazing fast Envpool Atari vectorized environment with EnvPool's XLA interface and JAX. |
| :material-github: [`ppo_atari_envpool_xla_jax_scan.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax_scan.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jax_scanpy) | Uses native `jax.scan` as opposed to python loops for faster compilation time. |
| :material-github: [`ppo_procgen.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_procgenpy) | For the procgen environments. |
| :material-github: [`ppo_atari_multigpu.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_multigpupy)| For Atari environments leveraging multi-GPUs. |
| :material-github: [`ppo_pettingzoo_ma_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_pettingzoo_ma_atari.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_pettingzoo_ma_ataripy)| For Pettingzoo's multi-agent Atari environments. |
Below are our single-file implementations of PPO:
## `ppo.py`
The [ppo.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py) has the following features:
* Works with the `Box` observation space of low-level features
* Works with the `Discrete` action space
* Works with envs like `CartPole-v1`
### Usage
=== "uv"
```bash
uv pip install .
uv run python cleanrl/ppo.py --help
uv run python cleanrl/ppo.py --env-id CartPole-v1
```
=== "pip"
```bash
python cleanrl/ppo.py --help
python cleanrl/ppo.py --env-id CartPole-v1
```
### Explanation of the logged metrics
Running `python cleanrl/ppo.py` will automatically record various metrics such as actor or value losses in Tensorboard. Below is the documentation for these metrics:
* `charts/episodic_return`: episodic return of the game
* `charts/episodic_length`: episodic length of the game
* `charts/SPS`: number of steps per second
* `charts/learning_rate`: the current learning rate
* `losses/value_loss`: the mean value loss across all data points
* `losses/policy_loss`: the mean policy loss across all data points
* `losses/entropy`: the mean entropy value across all data points
* `losses/old_approx_kl`: the approximate Kullback–Leibler divergence, measured by `(-logratio).mean()`, which corresponds to the k1 estimator in John Schulman’s blog post on [approximating KL](http://joschu.net/blog/kl-approx.html)
* `losses/approx_kl`: better alternative to `olad_approx_kl` measured by `(logratio.exp() - 1) - logratio`, which corresponds to the k3 estimator in [approximating KL](http://joschu.net/blog/kl-approx.html)
* `losses/clipfrac`: the fraction of the training data that triggered the clipped objective
* `losses/explained_variance`: the explained variance for the value function
### Implementation details
[ppo.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py) is based on the "13 core implementation details" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
1. Vectorized architecture (:material-github: [common/cmd_util.py#L22](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/cmd_util.py#L22))
1. Orthogonal Initialization of Weights and Constant Initialization of biases (:material-github: [a2c/utils.py#L58)](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L58))
1. The Adam Optimizer's Epsilon Parameter (:material-github: [ppo2/model.py#L100](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L100))
1. Adam Learning Rate Annealing (:material-github: [ppo2/ppo2.py#L133-L135](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/ppo2.py#L133-L135))
1. Generalized Advantage Estimation (:material-github: [ppo2/runner.py#L56-L65](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/runner.py#L56-L65))
1. Mini-batch Updates (:material-github: [ppo2/ppo2.py#L157-L166](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/ppo2.py#L157-L166))
1. Normalization of Advantages (:material-github: [ppo2/model.py#L139](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L139))
1. Clipped surrogate objective (:material-github: [ppo2/model.py#L81-L86](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L81-L86))
1. Value Function Loss Clipping (:material-github: [ppo2/model.py#L68-L75](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L68-L75))
1. Overall Loss and Entropy Bonus (:material-github: [ppo2/model.py#L91](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L91))
1. Global Gradient Clipping (:material-github: [ppo2/model.py#L102-L108](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L102-L108))
1. Debug variables (:material-github: [ppo2/model.py#L115-L116](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L115-L116))
1. Separate MLP networks for policy and value functions (:material-github: [common/policies.py#L156-L160](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/policies.py#L156-L160), [baselines/common/models.py#L75-L103](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L75-L103))
### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:3:8"
```
Below are the average episodic returns for `ppo.py`. To ensure the quality of the implementation, we compared the results against `openai/baselies`' PPO.
| Environment | `ppo.py` | `openai/baselies`' PPO (Huang et al., 2022)[^1]
| ----------- | ----------- | ----------- |
| CartPole-v1 | 490.04 ± 6.12 |497.54 ± 4.02 |
| Acrobot-v1 | -86.36 ± 1.32 | -81.82 ± 5.58 |
| MountainCar-v0 | -200.00 ± 0.00 | -200.00 ± 0.00 |
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh::9"
```
Tracked experiments and game play videos:
### Video tutorial
If you'd like to learn `ppo.py` in-depth, consider checking out the following video tutorial:
Tracked experiments and game play videos:
### Video tutorial
If you'd like to learn `ppo_atari.py` in-depth, consider checking out the following video tutorial:
## `ppo_continuous_action.py`
The [ppo_continuous_action.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py) has the following features:
* For continuous action space. Also implemented Mujoco-specific code-level optimizations
* Works with the `Box` observation space of low-level features
* Works with the `Box` (continuous) action space
* adding experimental support for [Gymnasium](https://gymnasium.farama.org/)
* 🧪 support `dm_control` environments via [Shimmy](https://github.com/Farama-Foundation/Shimmy)
???+ warning
We are now recommending users to use [`rpo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/rpo_continuous_action.py) instead of [`ppo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py) because `rpo_continuous_action.py` empirically performs better than `ppo_continuous_action.py` in 93% of the environments we tested. Please see [experiment results](/rl-algorithms/rpo/#experiment-results) for detailed analysis.
### Usage
=== "poetry"
```bash
# mujoco v4 environments
uv pip install ".[mujoco]"
python cleanrl/ppo_continuous_action.py --help
python cleanrl/ppo_continuous_action.py --env-id Hopper-v4
# dm_control environments
uv pip install ".[mujoco, dm_control]"
python cleanrl/ppo_continuous_action.py --env-id dm_control/cartpole-balance-v0
```
=== "pip"
```bash
pip install -r requirements/requirements-mujoco.txt
python cleanrl/ppo_continuous_action.py --help
python cleanrl/ppo_continuous_action.py --env-id Hopper-v4
pip install -r requirements/requirements-dm_control.txt
python cleanrl/ppo_continuous_action.py --env-id dm_control/cartpole-balance-v0
```
???+ warning "dm_control installation issue"
If you run into error like `AttributeError: 'GLFWContext' object has no attribute '_context'` in Linux, it's because the rendering dependencies are not installed properly. To fix it, try running
```
sudo apt-get update && sudo apt-get -y install libgl1-mesa-glx libosmesa6 libglfw3
```
See [https://github.com/deepmind/dm_control#rendering](https://github.com/deepmind/dm_control#rendering) for more detail.
### Explanation of the logged metrics
See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
### Implementation details
[ppo_continuous_action.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py) is based on the "9 details for continuous action domains (e.g. Mujoco)" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
1. Continuous actions via normal distributions (:material-github: [common/distributions.py#L103-L104](https://github.com/openai/baselines/blob/9b68103b737ac46bc201dfb3121cfa5df2127e53/baselines/common/distributions.py#L103-L104))
2. State-independent log standard deviation (:material-github: [common/distributions.py#L104](https://github.com/openai/baselines/blob/9b68103b737ac46bc201dfb3121cfa5df2127e53/baselines/common/distributions.py#L104))
3. Independent action components (:material-github: [common/distributions.py#L238-L246](https://github.com/openai/baselines/blob/9b68103b737ac46bc201dfb3121cfa5df2127e53/baselines/common/distributions.py#L238-L246))
4. Separate MLP networks for policy and value functions (:material-github: [common/policies.py#L160](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/policies.py#L160), [baselines/common/models.py#L75-L103](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L75-L103)
5. Handling of action clipping to valid range and storage (:material-github: [common/cmd_util.py#L99-L100](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/cmd_util.py#L99-L100))
6. Normalization of Observation (:material-github: [common/vec_env/vec_normalize.py#L4](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L4))
7. Observation Clipping (:material-github: [common/vec_env/vec_normalize.py#L39](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L39))
8. Reward Scaling (:material-github: [common/vec_env/vec_normalize.py#L28](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L28))
9. Reward Clipping (:material-github: [common/vec_env/vec_normalize.py#L32](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L32))
### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
MuJoCo v4
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:25:30"
```
{!benchmark/ppo_continuous_action.md!}
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh:11:19"
```
Tracked experiments and game play videos:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:36:41"
```
Below are the average episodic returns for `ppo_continuous_action.py` in `dm_control` environments.
| | ppo_continuous_action ({'tag': ['v1.0.0-13-gcbd83f6']}) |
|:--------------------------------------|:----------------------------------------------------------|
| dm_control/acrobot-swingup-v0 | 27.84 ± 9.25 |
| dm_control/acrobot-swingup_sparse-v0 | 1.60 ± 1.17 |
| dm_control/ball_in_cup-catch-v0 | 900.78 ± 5.26 |
| dm_control/cartpole-balance-v0 | 855.47 ± 22.06 |
| dm_control/cartpole-balance_sparse-v0 | 999.93 ± 0.10 |
| dm_control/cartpole-swingup-v0 | 640.86 ± 11.44 |
| dm_control/cartpole-swingup_sparse-v0 | 51.34 ± 58.35 |
| dm_control/cartpole-two_poles-v0 | 203.86 ± 11.84 |
| dm_control/cartpole-three_poles-v0 | 164.59 ± 3.23 |
| dm_control/cheetah-run-v0 | 432.56 ± 82.54 |
| dm_control/dog-stand-v0 | 307.79 ± 46.26 |
| dm_control/dog-walk-v0 | 120.05 ± 8.80 |
| dm_control/dog-trot-v0 | 76.56 ± 6.44 |
| dm_control/dog-run-v0 | 60.25 ± 1.33 |
| dm_control/dog-fetch-v0 | 34.26 ± 2.24 |
| dm_control/finger-spin-v0 | 590.49 ± 171.09 |
| dm_control/finger-turn_easy-v0 | 180.42 ± 44.91 |
| dm_control/finger-turn_hard-v0 | 61.40 ± 9.59 |
| dm_control/fish-upright-v0 | 516.21 ± 59.52 |
| dm_control/fish-swim-v0 | 87.91 ± 6.83 |
| dm_control/hopper-stand-v0 | 2.72 ± 1.72 |
| dm_control/hopper-hop-v0 | 0.52 ± 0.48 |
| dm_control/humanoid-stand-v0 | 6.59 ± 0.18 |
| dm_control/humanoid-walk-v0 | 1.73 ± 0.03 |
| dm_control/humanoid-run-v0 | 1.11 ± 0.04 |
| dm_control/humanoid-run_pure_state-v0 | 0.98 ± 0.03 |
| dm_control/humanoid_CMU-stand-v0 | 4.79 ± 0.18 |
| dm_control/humanoid_CMU-run-v0 | 0.88 ± 0.05 |
| dm_control/manipulator-bring_ball-v0 | 0.50 ± 0.29 |
| dm_control/manipulator-bring_peg-v0 | 1.80 ± 1.58 |
| dm_control/manipulator-insert_ball-v0 | 35.50 ± 13.04 |
| dm_control/manipulator-insert_peg-v0 | 60.40 ± 21.76 |
| dm_control/pendulum-swingup-v0 | 242.81 ± 245.95 |
| dm_control/point_mass-easy-v0 | 273.95 ± 362.28 |
| dm_control/point_mass-hard-v0 | 143.25 ± 38.12 |
| dm_control/quadruped-walk-v0 | 239.03 ± 66.17 |
| dm_control/quadruped-run-v0 | 180.44 ± 32.91 |
| dm_control/quadruped-escape-v0 | 28.92 ± 11.21 |
| dm_control/quadruped-fetch-v0 | 193.97 ± 22.20 |
| dm_control/reacher-easy-v0 | 626.28 ± 15.51 |
| dm_control/reacher-hard-v0 | 443.80 ± 9.64 |
| dm_control/stacker-stack_2-v0 | 75.68 ± 4.83 |
| dm_control/stacker-stack_4-v0 | 68.02 ± 4.02 |
| dm_control/swimmer-swimmer6-v0 | 158.19 ± 10.22 |
| dm_control/swimmer-swimmer15-v0 | 131.94 ± 0.88 |
| dm_control/walker-stand-v0 | 564.46 ± 235.22 |
| dm_control/walker-walk-v0 | 392.51 ± 56.25 |
| dm_control/walker-run-v0 | 125.92 ± 10.01 |
Note that the dm_control/lqr-lqr_2_1-v0 dm_control/lqr-lqr_6_2-v0 environments are never terminated or truncated. See https://wandb.ai/openrlbenchmark/cleanrl/runs/3tm00923 and https://wandb.ai/openrlbenchmark/cleanrl/runs/1z9us07j as an example.
Learning curves:

Tracked experiments and game play videos:
???+ info
In the gymnasium environments, we use the v4 mujoco environments, which roughly results in the same performance as the v2 mujoco environments.

### Video tutorial
If you'd like to learn `ppo_continuous_action.py` in-depth, consider checking out the following video tutorial:
## `ppo_atari_lstm.py`
The [ppo_atari_lstm.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py) has the following features:
* For Atari games using LSTM without stacked frames. It uses convolutional layers and common atari-based pre-processing techniques.
* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
* Works with the `Discrete` action space
### Usage
=== "poetry"
```bash
uv pip install ".[atari]"
uv run python cleanrl/ppo_atari_lstm.py --help
uv run python cleanrl/ppo_atari_lstm.py --env-id BreakoutNoFrameskip-v4
```
=== "pip"
```bash
pip install -r requirements/requirements-atari.txt
python cleanrl/ppo_atari_lstm.py --help
python cleanrl/ppo_atari_lstm.py --env-id BreakoutNoFrameskip-v4
```
### Explanation of the logged metrics
See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
### Implementation details
[ppo_atari_lstm.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py) is based on the "5 LSTM implementation details" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
1. Layer initialization for LSTM layers (:material-github: [a2c/utils.py#L84-L86](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L84-L86))
2. Initialize the LSTM states to be zeros (:material-github: [common/models.py#L179](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L179))
3. Reset LSTM states at the end of the episode (:material-github: [common/models.py#L141](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L141))
4. Prepare sequential rollouts in mini-batches (:material-github: [a2c/utils.py#L81](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L81))
5. Reconstruct LSTM states during training (:material-github: [a2c/utils.py#L81](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L81))
To help test out the memory, we remove the 4 stacked frames from the observation (i.e., using `env = gym.wrappers.FrameStack(env, 1)` instead of `env = gym.wrappers.FrameStack(env, 4)` like in `ppo_atari.py` )
### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:47:52"
```
Below are the average episodic returns for `ppo_atari_lstm.py`. To ensure the quality of the implementation, we compared the results against `openai/baselies`' PPO.
| Environment | `ppo_atari_lstm.py` | `openai/baselies`' PPO (Huang et al., 2022)[^1]
| ----------- | ----------- | ----------- |
| BreakoutNoFrameskip-v4 | 128.92 ± 31.10 | 138.98 ± 50.76 |
| PongNoFrameskip-v4 | 19.78 ± 1.58 | 19.79 ± 0.67 |
| BeamRiderNoFrameskip-v4 | 1536.20 ± 612.21 | 1591.68 ± 372.95|
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh:11:19"
```
Tracked experiments and game play videos:
## `ppo_atari_envpool.py`
The [ppo_atari_envpool.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py) has the following features:
* Uses the blazing fast [Envpool](https://github.com/sail-sg/envpool) vectorized environment.
* For Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
* Works with the `Discrete` action space
???+ warning
Note that `ppo_atari_envpool.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. See envpool's built wheels here: [https://pypi.org/project/envpool/#files](https://pypi.org/project/envpool/#files)
???+ bug
EnvPool's vectorized environment **does not behave the same** as gym's vectorized environment, which causes a compatibility bug in our PPO implementation. When an action $a$ results in an episode termination or truncation, the environment generates $s_{last}$ as the terminated or truncated state; we then use $s_{new}$ to denote the initial state of the new episodes. Here is how the bahviors differ:
* Under the vectorized environment of `envpool<=0.6.4`, the `obs` in `obs, reward, done, info = env.step(action)` is the truncated state $s_{last}$
* Under the vectorized environment of `gym==0.23.1`, the `obs` in `obs, reward, done, info = env.step(action)` is the initial state $s_{new}$.
This causes the $s_{last}$ to be off by one.
See [:material-github: sail-sg/envpool#194](https://github.com/sail-sg/envpool/issues/194) for more detail. However, it does not seem to impact performance, so we take a note here and await for the upstream fix.
### Usage
=== "poetry"
```bash
uv pip install ".[envpool]"
uv run python cleanrl/ppo_atari_envpool.py --help
uv run python cleanrl/ppo_atari_envpool.py --env-id Breakout-v5
```
=== "pip"
```bash
pip install -r requirements/requirements-envpool.txt
python cleanrl/ppo_atari_envpool.py --help
python cleanrl/ppo_atari_envpool.py --env-id Breakout-v5
```
### Explanation of the logged metrics
See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
### Implementation details
[ppo_atari_envpool.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py) uses a customized `RecordEpisodeStatistics` to work with envpool but has the same other implementation details as `ppo_atari.py` (see [related docs](/rl-algorithms/ppo/#implementation-details_1)).
### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:58:63"
```
{!benchmark/ppo_atari_envpool.md!}
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh:51:62"
```
Tracked experiments and game play videos:
## `ppo_atari_envpool_xla_jax.py`
The [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) has the following features:
* Uses the blazing fast [Envpool](https://github.com/sail-sg/envpool) vectorized environment.
* Uses EnvPool's experimental [XLA interface](https://envpool.readthedocs.io/en/latest/content/xla_interface.html).
* Uses [Jax](https://github.com/google/jax), [Flax](https://github.com/google/flax), and [Optax](https://github.com/deepmind/optax) instead of `torch`.
* For Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
* Works with the `Discrete` action space
???+ warning
Note that `ppo_atari_envpool_xla_jax.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. See envpool's built wheels here: [https://pypi.org/project/envpool/#files](https://pypi.org/project/envpool/#files)
???+ bug
EnvPool's vectorized environment **does not behave the same** as gym's vectorized environment, which causes a compatibility bug in our PPO implementation. When an action $a$ results in an episode termination or truncation, the environment generates $s_{last}$ as the terminated or truncated state; we then use $s_{new}$ to denote the initial state of the new episodes. Here is how the bahviors differ:
* Under the vectorized environment of `envpool<=0.6.4`, the `obs` in `obs, reward, done, info = env.step(action)` is the truncated state $s_{last}$
* Under the vectorized environment of `gym==0.23.1`, the `obs` in `obs, reward, done, info = env.step(action)` is the initial state $s_{new}$.
This causes the $s_{last}$ to be off by one.
See [:material-github: sail-sg/envpool#194](https://github.com/sail-sg/envpool/issues/194) for more detail. However, it does not seem to impact performance, so we take a note here and await for the upstream fix.
### Usage
=== "poetry"
```bash
uv pip install ".[envpool, jax]"
uv pip install --upgrade "jax[cuda11_cudnn82]==0.4.8" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
uv run python cleanrl/ppo_atari_envpool_xla_jax.py --help
uv run python cleanrl/ppo_atari_envpool_xla_jax.py --env-id Breakout-v5
```
=== "pip"
```bash
pip install -r requirements/requirements-envpool.txt
pip install -r requirements/requirements-jax.txt
pip install --upgrade "jax[cuda]==0.3.17" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
python cleanrl/ppo_atari_envpool_xla_jax.py --help
python cleanrl/ppo_atari_envpool_xla_jax.py --env-id Breakout-v5
```
### Explanation of the logged metrics
See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`. In [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) we omit logging `losses/old_approx_kl` and `losses/clipfrac` for brevity.
Additionally, we record the following metric:
* `charts/avg_episodic_return`: the average value of the *latest* episodic returns of `args.num_envs=8` envs
* `charts/avg_episodic_length`: the average value of the *latest* episodic lengths of `args.num_envs=8` envs
???+ info
Note that we use `charts/avg_episodic_return` and `charts/avg_episodic_length` in place of `charts/episodic_return` and `charts/episodic_length` because under the EnvPool's XLA interface, we can only record fixed-shape metrics where as there could be a variable number of raw episodic returns / lengths. To resolve this challenge, we create variables (e.g., `returned_episode_returns`, `returned_episode_lengths`) to keep track of the *latest* episodic returns / lengths of each environment and average them for reporting purposes.
### Implementation details
[ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) uses the same other implementation details as `ppo_atari.py` (see [related docs](/rl-algorithms/ppo/#implementation-details_1)), with two differences
1. [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) does not use the value function clipping by default, because there is no sufficient evidence that value function clipping actually improves performance.
1. [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) uses a customized `EpisodeStatistics` to record episode statistics instead of the `RecordEpisodeStatistics` used in other variants. `RecordEpisodeStatistics` is a *stateful* python wrapper which is incompatible with EnvPool's *stateless* XLA interface. To address this issue, we used a `EpisodeStatistics` dataclass and simply implement the logic of `RecordEpisodeStatistics`. However, `EpisodeStatistics` comes with a major limitation: its storage has a fixed shape and can only record the *latest* episodic return of the sub-environments. Furthermore, the default episodic return values in `EpisodeStatistics` are set to zeros, which does not necessarily correspond to the episodic return obtained by a random policy. For example, we would report `charts/avg_episodic_return=0` for `Pong-v5`, even if they should have been `charts/avg_episodic_return=-21`. That said, this issue goes away as soon as the sub-environments finished their first episodes, therefore not impacting the reported results.
???+ info
We benchmarked the PPO implementation w/ and w/o value function clipping, finding no significant difference in performance, which is consistent with the findings in Andrychowicz et al.[^2]. See the related report [part 1](https://wandb.ai/costa-huang/cleanRL/reports/CleanRL-PPO-JAX-EnvPool-s-XLA-w-and-w-o-value-loss-clipping-vs-openai-baselins-PPO-part-1---VmlldzoyNzQ3MzQ1) and [part 2](https://wandb.ai/costa-huang/cleanRL/reports/CleanRL-PPO-JAX-EnvPool-s-XLA-w-and-w-o-value-loss-clipping-vs-openai-baselins-PPO-part-2---VmlldzoyNzQ3MzUw).

### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:69:74"
```
{!benchmark/ppo_atari_envpool_xla_jax.md!}
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh:64:85"
```
???+ info
Note the original openai/baselines uses `atari-py==0.2.6` which hangs on `gym.make("DefenderNoFrameskip-v4")` and does not support SurroundNoFrameskip-v4 (see issue [:material-github: openai/atari-py#73](https://github.com/openai/atari-py/issues/73)). To get results on these environments, we use `gym==0.23.1 ale-py==0.7.4 "AutoROM[accept-rom-license]==0.4.2` and [manually register `SurroundNoFrameskip-v4` in our fork](https://github.com/vwxyzjn/baselines/blob/e2cb1c938a62fa8d7fe98187246cde08dfd57bd1/baselines/common/register_all_atari_envs.py#L2).
Median Human Normalized Score (HNS) compared to SEEDRL's R2D2 (data available [here](https://github.com/google-research/seed_rl/blob/66e8890261f09d0355e8bf5f1c5e41968ca9f02b/docs/seed_r2d2_atari_graphs.csv)).

???+ info
Note the SEEDRL's R2D2's median HNS data does not include learning curves for `Defender` and `Surround` (see [google-research/seed_rl#78](https://github.com/google-research/seed_rl/issues/78)). Also note the SEEDRL's R2D2 uses slightly different Atari preprocessing than our `ppo_atari_envpool_xla_jax.py`, so we may be comparing apples and oranges; however, the results are still informative at the scale of 57 Atari games — we would be at least comparing similar apples.
Tracked experiments and game play videos:
## `ppo_atari_envpool_xla_jax_scan.py`
The [ppo_atari_envpool_xla_jax_scan.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax_scan.py) has the following features:
* Replaces python loops in `compute_gae`, `update_ppo`, and `rollout` functions of [ppo_atari_envpool_xla_jax.py](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy) with native `jax.scan`
* Warnings and caveats from [ppo_atari_envpool_xla_jax.py](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy) also apply here
### Usage
=== "poetry"
```bash
uv pip install ".[envpool, jax]"
uv pip install --upgrade "jax[cuda11_cudnn82]==0.4.8" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
uv run python cleanrl/ppo_atari_envpool_xla_jax_scan.py --help
uv run python cleanrl/ppo_atari_envpool_xla_jax_scan.py --env-id Breakout-v5
```
=== "pip"
```bash
pip install -r requirements/requirements-envpool.txt
pip install -r requirements/requirements-jax.txt
pip install --upgrade "jax[cuda]==0.3.17" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
python cleanrl/ppo_atari_envpool_xla_jax_scan.py --help
python cleanrl/ppo_atari_envpool_xla_jax_scan.py --env-id Breakout-v5
```
### Explanation of the logged metrics
See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`. The metrics are the same as those in [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py).
### Implementation details
[ppo_atari_envpool_xla_jax_scan.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax_scan.py) is a clone of [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) that replaces the python loops with native `jax.scan`.
### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:80:85"
```
{!benchmark/ppo_atari_envpool_xla_jax_scan.md!}
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh:87:96"
```
Learning curves:
???+ info
The training time of this variant and that of [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) are very similar but the compilation time is reduced significantly (see [vwxyzjn/cleanrl#328](https://github.com/vwxyzjn/cleanrl/pull/328#issuecomment-1340474894)). Note that the hardware also affects the speed in the learning curve below. Runs from [`costa-huang`](https://github.com/vwxyzjn/) (red) are slower from those of [`51616`](https://github.com/51616/) (blue and orange) because of hardware differences.


Tracked experiments:
## `ppo_procgen.py`
The [ppo_procgen.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py) has the following features:
* For the procgen environments
* Uses IMPALA-style neural network
* Works with the `Discrete` action space
### Usage
=== "poetry"
```bash
uv pip install ".[procgen]"
uv run python cleanrl/ppo_procgen.py --help
uv run python cleanrl/ppo_procgen.py --env-id starpilot
```
=== "pip"
```bash
pip install -r requirements/requirements-procgen.txt
python cleanrl/ppo_procgen.py --help
python cleanrl/ppo_procgen.py --env-id starpilot
```
### Explanation of the logged metrics
See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
### Implementation details
[ppo_procgen.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py) is based on the details in "Appendix" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
1. IMPALA-style Neural Network (:material-github: [common/models.py#L28](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L28))
1. Use the same `gamma` parameter in the `NormalizeReward` wrapper. Note that the original implementation from [openai/train-procgen](https://github.com/openai/train-procgen) uses the default `gamma=0.99` in [the `VecNormalize` wrapper](https://github.com/openai/train-procgen/blob/1a2ae2194a61f76a733a39339530401c024c3ad8/train_procgen/train.py#L43) but `gamma=0.999` as PPO's parameter. The mismatch between the `gamma`s is technically incorrect. See [#209](https://github.com/vwxyzjn/cleanrl/pull/209)
### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:91:100"
```
We try to match the default setting in [openai/train-procgen](https://github.com/openai/train-procgen) except that we use the `easy` distribution mode and `total_timesteps=25e6` to save compute. Notice [openai/train-procgen](https://github.com/openai/train-procgen) has the following settings:
1. Learning rate annealing is turned off by default
1. Reward scaling and reward clipping is used
Below are the average episodic returns for `ppo_procgen.py`. To ensure the quality of the implementation, we compared the results against `openai/baselies`' PPO.
| Environment | `ppo_procgen.py` | `openai/baselies`' PPO (Huang et al., 2022)[^1]
| ----------- | ----------- | ----------- |
| StarPilot (easy) | 30.99 ± 1.96 | 33.97 ± 7.86 |
| BossFight (easy) | 8.85 ± 0.33 | 9.35 ± 2.04 |
| BigFish (easy) | 16.46 ± 2.71 | 20.06 ± 5.34 |
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh:98:106"
```
???+ info
Note that we have run the procgen experiments using the `easy` distribution for reducing the computational cost.
Tracked experiments and game play videos:
## `ppo_atari_multigpu.py`
The [ppo_atari_multigpu.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py) leverages data parallelism to speed up training time *at no cost of sample efficiency*.
`ppo_atari_multigpu.py` has the following features:
* Allows the users to use do training leveraging data parallelism
* For playing Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
* Works with the `Discrete` action space
???+ warning
Note that `ppo_atari_multigpu.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. It will error out with `NOTE: Redirects are currently not supported in Windows or MacOs.` See [pytorch/pytorch#20380](https://github.com/pytorch/pytorch/issues/20380)
### Usage
=== "poetry"
```bash
uv pip install ".[atari]"
uv run python cleanrl/ppo_atari_multigpu.py --help
# `--nproc_per_node=2` specifies how many subprocesses we spawn for training with data parallelism
# note it is possible to run this with a *single GPU*: each process will simply share the same GPU
uv run torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4
# by default we use the `gloo` backend, but you can use the `nccl` backend for better multi-GPU performance
uv run torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --backend nccl
# it is possible to spawn more processes than the amount of GPUs you have via `--device-ids`
# e.g., the command below spawns two processes using GPU 0 and two processes using GPU 1
uv run torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --device-ids 0 0 1 1
```
=== "pip"
```bash
pip install -r requirements/requirements-atari.txt
python cleanrl/ppo_atari_multigpu.py --help
# `--nproc_per_node=2` specifies how many subprocesses we spawn for training with data parallelism
# note it is possible to run this with a *single GPU*: each process will simply share the same GPU
torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4
# by default we use the `gloo` backend, but you can use the `nccl` backend for better multi-GPU performance
torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --backend nccl
# it is possible to spawn more processes than the amount of GPUs you have via `--device-ids`
# e.g., the command below spawns two processes using GPU 0 and two processes using GPU 1
torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --device-ids 0 0 1 1
```
### Explanation of the logged metrics
See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
### Implementation details
[ppo_atari_multigpu.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py) is based on `ppo_atari.py` (see its [related docs](/rl-algorithms/ppo/#implementation-details_1)).
We use [Pytorch's distributed API](https://pytorch.org/tutorials/intermediate/dist_tuto.html) to implement the data parallelism paradigm. The basic idea is that the user can spawn $N$ processes each running a copy of `ppo_atari.py`, holding a copy of the model, stepping the environments, and averaging their gradients together for the backward pass. Here are a few note-worthy implementation details.
1. **Local versus global parameters**: All of the parameters in `ppo_atari.py` are global (such as batch size), but in `ppo_atari_multigpu.py` we have local parameters as well. Say we run `torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --local-num-envs=4`; here are how all multi-gpu related parameters are adjusted:
* **number of environments**: `num_envs = local_num_envs * world_size = 4 * 2 = 8`
* **batch size**: `local_batch_size = local_num_envs * num_steps = 4 * 128 = 512`, `batch_size = num_envs * num_steps) = 8 * 128 = 1024`
* **minibatch size**: `local_minibatch_size = int(args.local_batch_size // args.num_minibatches) = 512 // 4 = 128`, `minibatch_size = int(args.batch_size // args.num_minibatches) = 1024 // 4 = 256`
* **number of updates**: `num_iterations = args.total_timesteps // args.batch_size = 10000000 // 1024 = 9765`
1. **Adjust seed per process**: we need be very careful with seeding: we could have used the exact same seed for each subprocess. To ensure this does not happen, we do the following
```python hl_lines="2 5 16"
# CRUCIAL: note that we needed to pass a different seed for each data parallelism worker
args.seed += local_rank
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed - local_rank)
torch.backends.cudnn.deterministic = args.torch_deterministic
# ...
envs = gym.vector.SyncVectorEnv(
[make_env(args.env_id, args.seed + i, i, args.capture_video, run_name) for i in range(args.num_envs)]
)
assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
agent = Agent(envs).to(device)
torch.manual_seed(args.seed)
optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
```
Notice that we adjust the seed with `args.seed += local_rank` (line 2), where `local_rank` is the index of the subprocesses. This ensures we seed packages and envs with uncorrealted seeds. However, we do need to use the same `torch` seed for all process to initialize same weights for the `agent` (line 5), after which we can use a different seed for `torch` (line 16).
1. **Efficient gradient averaging**: PyTorch recommends to average the gradient across the whole world via the following (see [docs](https://pytorch.org/tutorials/intermediate/dist_tuto.html#distributed-training))
```python
for param in agent.parameters():
dist.all_reduce(param.grad.data, op=dist.ReduceOp.SUM)
param.grad.data /= world_size
```
However, [@cswinter](https://github.com/cswinter) introduces a more efficient gradient averaging scheme with proper batching (see :material-github: [entity-neural-network/incubator#220](https://github.com/entity-neural-network/incubator/pull/220)), which looks like:
```python
all_grads_list = []
for param in agent.parameters():
if param.grad is not None:
all_grads_list.append(param.grad.view(-1))
all_grads = torch.cat(all_grads_list)
dist.all_reduce(all_grads, op=dist.ReduceOp.SUM)
offset = 0
for param in agent.parameters():
if param.grad is not None:
param.grad.data.copy_(
all_grads[offset : offset + param.numel()].view_as(param.grad.data) / world_size
)
offset += param.numel()
```
In our previous empirical testing (see :material-github: [vwxyzjn/cleanrl#162](https://github.com/vwxyzjn/cleanrl/pull/162#issuecomment-1107909696)), we have found [@cswinter](https://github.com/cswinter)'s implementation to be faster, hence we adopt it in our implementation.
### Experiment results
To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
``` title="benchmark/ppo.sh" linenums="1"
--8<-- "benchmark/ppo.sh:102:107"
```
Below are the average episodic returns for `ppo_atari_multigpu.py`. To ensure no loss of sample efficiency, we compared the results against `ppo_atari.py`.
{!benchmark/ppo_atari_multigpu.md!}
Learning curves:
``` title="benchmark/ppo_plot.sh" linenums="1"
--8<-- "benchmark/ppo_plot.sh:108:117"
```
Under the same hardware, we see that `ppo_atari_multigpu.py` is about **30% faster** than `ppo_atari.py` with no loss of sample efficiency.
???+ info
The experiments above is to show correctness -- we show that by aligning the same hyperparameters of `ppo_atari.py` and `ppo_atari_multigpu.py`, we can achieve the same sample efficiency. However, we can train even faster by simply running a much larger batch size. For example, we can run `torchrun --standalone --nnodes=1 --nproc_per_node=8 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --local-num-envs=8`, which will run 8 x 8 = 64 environments in parallel and achieve a batch size of 64 x 128 = 8192. This will likely result in a sample efficiency but should increase the wall time efficiency.
???+ info
Although `ppo_atari_multigpu.py` is 30% faster than `ppo_atari.py`, `ppo_atari_multigpu.py` is still slower than `ppo_atari_envpool.py`, as shown below. This comparison really highlights the different kinds of optimization possible.