Quick Start Guide

Installation & Setup

1. Install Dependencies

First, install the required Python packages:

2. Configure Environment Variables

Set up your API credentials by exporting the following environment variables:

3. Start the Server

Launch the AgentCity server using uvicorn:

Verify Migration

To verify that the models have been successfully migrated and are working correctly:

1. Download LibCity datasets: Download the required datasets and place them in the /Bigscity-LibCity/raw_data directory.
2. Run a model: Execute the following command to test a migrated model:
   python run_model.py --task traffic_state_pred --model model_name --dataset dataset_name
3. Replace placeholders:
   • model_name: One of the migrated models (e.g., PatchSTG, LSTGAN, MLCAFormer)
   • dataset_name: Your target dataset name
4. Check results: After execution, check the output for evaluation metrics and verify the model is running correctly.

Next Steps

Now that you have AgentCity set up, you can:

• Explore the LeaderBoard to see performance metrics of migrated models
• Experiment with different models and datasets
• Contribute new model migrations to the project
• Tune hyperparameters to improve model performance