### Install Core EPyMARL Dependencies

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Installs the main dependencies for the EPyMARL framework. Run this after cloning the repository.

```sh
pip install -r requirements.txt
```

--------------------------------

### Install Supported Environment Dependencies

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Installs a set of supported multi-agent environments for EPyMARL. This includes various MARL benchmarks and simulators.

```sh
pip install -r env_requirements.txt
```

--------------------------------

### Run MAPPO with Individual Rewards in LBF

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Example command to run the MAPPO algorithm in the Level-Based Foraging environment with individual rewards enabled. Ensure the environment and algorithm support general-sum rewards.

```sh
python src/main.py --config=mappo --env-config=gymma with env_args.time_limit=50 env_args.key="lbforaging:Foraging-8x8-2p-3f-v3" common_reward=False
```

--------------------------------

### Configure Reward Scalarization to Mean

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Example configuration to change the reward scalarization method from the default sum to mean when using common rewards in environments that provide individual rewards.

```sh
python src/main.py --config=mappo --env-config=gymma with env_args.time_limit=50 env_args.key="lbforaging:Foraging-8x8-2p-3f-v3" common_reward=True reward_scalarisation="mean"
```

--------------------------------

### Run Pareto-AC Experiment in Penalty Game

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Example command to run the Pareto-AC algorithm in the Penalty game environment. It specifies the configuration, environment, and game parameters.

```sh
python src/main.py --config=pac_ns --env-config=gymma with env_args.time_limit=1 env_args.key=matrixgames:penalty-100-nostate-v0
```

--------------------------------

### Install Pareto-AC Algorithm Dependencies

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Installs the additional dependencies required for the Pareto Actor-Critic (PAC) algorithm. This should be run if you intend to use PAC.

```sh
pip install -r pac_requirements.txt
```

--------------------------------

### Run RWARE Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in the RWARE environment with a time limit of 500 and a specific map key.

```sh
python src/main.py --config=qmix --env-config=gymma with env_args.time_limit=500 env_args.key="rware:rware-tiny-2ag-v2"
```

--------------------------------

### Run Experiment in Custom Gymnasium Environment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute an experiment in a custom registered Gymnasium environment. Uses the gymma environment config and specifies the custom environment key.

```sh
python src/main.py --config=qmix --env-config=gymma with env_args.time_limit=50 env_args.key="myenv:my-environment-v1"
```

--------------------------------

### Run VMAS Environment Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute an experiment using a VMAS environment with the qmix algorithm. Sets a specific time limit and environment key.

```sh
python src/main.py --config=qmix --env-config=gymma with env_args.time_limit=150 env_args.key="vmas-balance"
```

--------------------------------

### Run SMAClite Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in the SMAClite environment with a time limit of 150 and a specific map name.

```sh
python src/main.py --config=qmix --env-config=smaclite with env_args.time_limit=150 env_args.map_name="MMM"
```

--------------------------------

### Configure Online Weights and Biases Logging

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

To directly log runs online to W&B, specify 'wandb_mode="online"' in the configuration.

```yaml
wandb_mode:"online"
```

--------------------------------

### Run Level-Based Foraging Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in the Level-Based Foraging environment with a time limit of 50 and a specific map key.

```sh
python src/main.py --config=qmix --env-config=gymma with env_args.time_limit=50 env_args.key="lbforaging:Foraging-8x8-2p-3f-v3"
```

--------------------------------

### Run Matrix Games Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in the Matrix Games environment with a time limit of 25 and a specific penalty configuration.

```sh
python src/main.py --config=qmix --env-config=gymma with env_args.time_limit=25 env_args.key="matrixgames:penalty-100-nostate-v0"
```

--------------------------------

### Run MPE Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in the MPE (Multi-Agent Particle Environments) with a time limit of 25 and a specific map key.

```sh
python src/main.py --config=qmix --env-config=gymma with env_args.time_limit=25 env_args.key="pz-mpe-simple-spread-v3"
```

--------------------------------

### Run SMAC Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in the SMAC (StarCraft Multi-Agent Challenge) environment with a specific map name.

```sh
python src/main.py --config=qmix --env-config=sc2 with env_args.map_name="3s5z"
```

--------------------------------

### Configure Weights and Biases Logging

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Set these parameters in your configuration to enable logging to Weights and Biases. By default, runs are logged in offline mode.

```yaml
use_wandb: True # Log results to W&B
wandb_team: null # W&B team name
wandb_project: null # W&B project name
```

--------------------------------

### Run SMACv2 Experiment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in the SMACv2 environment with a specific map name.

```sh
python src/main.py --config=qmix --env-config=sc2v2 with env_args.map_name="protoss_5_vs_5"
```

--------------------------------

### Run SMAClite Experiment with C++ RVO2

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a QMIX experiment in SMAClite, utilizing the C++ RVO2 library for collision avoidance.

```sh
python src/main.py --config=qmix --env-config=smaclite with env_args.time_limit=150 env_args.map_name="MMM" env_args.use_cpp_rvo2=True
```

--------------------------------

### Register Custom Gymnasium Environment

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Template for registering a custom environment with Gymnasium. Specify the environment ID, entry point, and initialization arguments.

```python
from gymnasium import register

register(
  id="my-environment-v1",                         # Environment ID.
  entry_point="myenv.environment:MyEnvironment",  # The entry point for the environment class
  kwargs={
            ...                                   # Arguments that go to MyEnvironment's __init__ function.
        },
    )
```

--------------------------------

### Run Hyperparameter Search Locally

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a hyperparameter search using a specified configuration file and number of seeds. Runs the search locally.

```sh
python search.py run --config=search.config.example.yaml --seeds 5 locally
```

--------------------------------

### Run Hyperparameter Search in Batch Script

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Execute a hyperparameter search for a single process within a cluster environment. Specify the configuration file, seeds, and the index of the specific combination.

```sh
python search.py run --config=search.config.example.yaml --seeds 5 single 1
```

--------------------------------

### Configure Weights and Biases Model Saving

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Enable saving models to W&B by setting 'wandb_save_model: True'. This requires 'use_wandb' and 'save_model' to also be true. Model saving frequency can be adjusted with 'save_model_interval'.

```yaml
wandb_save_model: True # Save models to W&B (only done if use_wandb is True and save_model is True)
save_model: True # Save the models to disk
save_model_interval: 50000
```

--------------------------------

### PyMARL Citation (BibTeX)

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Use this BibTeX entry to cite the original PyMARL codebase in academic publications. It includes title, authors, journal, volume, and year.

```tex
@article{samvelyan19smac,
  title = {{The} {StarCraft} {Multi}-{Agent} {Challenge}},
  author = {Mikayel Samvelyan and Tabish Rashid and Christian Schroeder de Witt and Gregory Farquhar and Nantas Nardelli and Tim G. J. Rudner and Chia-Man Hung and Philiph H. S. Torr and Jakob Foerster and Shimon Whiteson},
  journal = {CoRR},
  volume = {abs/1902.04043},
  year = {2019},
}
```

--------------------------------

### EPyMARL Citation (BibTeX)

Source: https://github.com/uoe-agents/epymarl/blob/main/README.md

Use this BibTeX entry to cite the EPyMARL codebase in academic publications. It includes title, authors, book title, year, and URLs for the paper and code.

```tex
@inproceedings{papoudakis2021benchmarking,
   title={Benchmarking Multi-Agent Deep Reinforcement Learning Algorithms in Cooperative Tasks},
   author={Georgios Papoudakis and Filippos Christianos and Lukas Schäfer and Stefano V. Albrecht},
   booktitle = {Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks (NeurIPS)},
   year={2021},
   url = {http://arxiv.org/abs/2006.07869},
   openreview = {https://openreview.net/forum?id=cIrPX-Sn5n},
   code = {https://github.com/uoe-agents/epymarl},
}
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.