MTrainS: Improving DLRM training efficiency using heterogeneous memories — Quantapedia

Recommendation models are very large, requiring terabytes (TB) of memory during training. In pursuit of better quality, the model size and complexity grow over time, which requires additional training