ml-agents is an open-source toolkit that enables games and simulations made with Unity to serve as environments for training intelligent agents.
Exploring ML-Agents in the Hub
You can find ml-agents models by filtering at the left of the models page.
All models on the Hub come up with useful features:
An automatically generated model card with a description, a training configuration, and more.
Metadata tags that help for discoverability.
Tensorboard summary files to visualize the training metrics.
A link to the Spaces web demo where you can visualize your agent playing in your browser.
Install the library
To install the ml-agents library, you need to clone the repo:
Copied
# Clone the repository
git clone https://github.com/Unity-Technologies/ml-agents
# Go inside the repository and install the package
cd ml-agents
pip3 install -e ./ml-agents-envs
pip3 install -e ./ml-agents
Using existing models
You can simply download a model from the Hub using mlagents-load-from-hf.
Copied
You need to define two parameters:
--repo-id: the name of the BOINC AI repo you want to download.
--local-dir: the path to download the model.
Visualize an agent playing
You can easily watch any model playing directly in your browser:
Go to your model repo.
In the Watch Your Agent Play section, click on the link.
In the demo, on step 1, choose your model repository, which is the model id.
In step 2, choose what model you want to replay.
Sharing your models
You can easily upload your models using mlagents-push-to-hf:
Copied
You need to define four parameters:
--run-id: the name of the training run id.
--local-dir: where the model was saved.
--repo-id: the name of the BOINC AI repo you want to create or update. It’s <your huggingface username>/<the repo name>.
