HoTPP Benchmark v0.1.0

1. Add extra fields beyond timestamps and labels.
2. Add transactions from Multimodal Banking Dataset (https://arxiv.org/abs/2409.17587).
3. Add embeddings computation script (hotpp.embed).

Contributors

@ivan-chai

HoTPP Benchmark v0.0.2

Minor changes:

• Add Max Accuracy metric
• Fix multi-gpu training
• Fix datasets with very short sequences
• Add MAE loss

Contributors

@ivan-chai

HoTPP Benchmark v0.0.1

The initial release of HoTPP Benchmark for the long-horizon event sequences forecasting.

Contributors

The core library and scripts are made by @ivan-chai. Thanks to @xiyori for providing MIMIC-IV preprocessing tools.