End-to-end session-based recommendation

These end-to-end example notebooks focus on the following:

• Preprocessing the Yoochoose e-commerce dataset.

• Generating session features with on GPU.

• Using the NVTabular dataloader with PyTorch.

• Training a session-based recommendation model with a Transformer architecture (XLNET).

• Exporting the preprocessing workflow and trained model to Triton Inference Server (TIS).

• Sending request to TIS and generating next-item predictions for each session.

Refer to the following notebooks:

• ETL with NVTabular

• End-to-end session-based recommendation: PyTorch