Migration of T-GCN to TensorFlow 2.x and PyTorch 2.x

We are thrilled to announce the successful migration of the Temporal Graph Convolutional Network (T-GCN) model to TensorFlow 2.x and PyTorch 2.x! This effort was led by Mr. Amiri & Mr. Mostafazade, members of the GTA-Lab, who have updated the original implementations to leverage the latest features and optimizations in these frameworks.

The T-GCN model, originally proposed in the paper T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction, combines Graph Convolutional Networks (GCN) and Gated Recurrent Units (GRU) to capture spatial and temporal dependencies in traffic data. The updated implementations ensure better compatibility, performance, and ease of use for researchers and practitioners.

Key Updates:

1. TensorFlow 2.x Implementation:
   The original TensorFlow 1.x code has been migrated to TensorFlow 2.x. The updated repository can be found here:
   https://github.com/GTA-Lab/TGCN-TF2.

2. PyTorch 2.x Implementation:
   The original PyTorch Lightning code has been migrated to PyTorch 2.x. The updated repository can be found here:
   https://github.com/GTA-Lab/TGCN-PyTorch2.

We extend our gratitude to Amiri & Mostafazade for their dedication and effort in making these updates available to the community. These repositories will serve as valuable resources for researchers working on traffic prediction and graph-based deep learning models.