witgaw/DCRNN_METR-LA

DCRNN Model - METR-LA

Diffusion Convolutional Recurrent Neural Network (DCRNN) trained on METR-LA dataset for traffic speed forecasting.

Model Description

This model uses a graph neural network architecture that combines:

• Diffusion convolution to capture spatial dependencies on road networks
• Recurrent neural networks (GRU) for temporal modeling
• Sequence-to-sequence learning for multi-step ahead forecasting

Evaluation Metrics

• MAE (val split): 2.7835

Dataset

METR-LA: Traffic speed data from highway sensors.

Usage

from dcrnn_pytorch.model.pytorch.dcrnn_supervisor import DCRNNSupervisor
from utils.dcrnn import load_from_hub

# Download checkpoint
checkpoint_path = load_from_hub("your-username/dcrnn-metr-la")

# Load model (requires config and adj_mx from checkpoint)
# See documentation for full loading example

Training

Model was trained using the DCRNN PyTorch implementation with default hyperparameters.

Citation

If you use this model, please cite the original DCRNN paper:

@inproceedings{li2018dcrnn,
  title={Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting},
  author={Li, Yaguang and Yu, Rose and Shahabi, Cyrus and Liu, Yan},
  booktitle={International Conference on Learning Representations (ICLR)},
  year={2018}
}

License

This model checkpoint is released under the same license as the training code.