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 the 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:
End-to-end session-based recommendation: TensorFlow | PyTorch