Add-to-Cart Prediction on Industrial-scale Recommender Dataset
87.36AUCSORT(Transformer+ST+LA+QP+AG+QKN+MoE)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SORT(Transformer+ST+LA+QP+AG+QKN+MoE)Scale=Large, Params=685M, FLOPs=188G2026.03 | 87.36 | 0.41 | |
| TransformerScale=Large, Params=144M, FLOPs=322G2026.03 | 86.95 | — | |
| SORT(Transformer+ST+LA+QP+AG+QKN+MoE)Scale=Base, Params=85M, FLOPs=24G2026.03 | 86.94 | 0.31 | |
| Transformer+special token (ST)Scale=Base, Params=18M, FLOPs=43G2026.03 | 86.85 | 0.22 | |
| OneTransScale=Base, Params=54M, FLOPs=24G2026.03 | 86.77 | 0.14 | |
| Transformer+query pruning (QP)Scale=Base, Params=18M, FLOPs=24G2026.03 | 86.76 | 0.13 | |
| Transformer+MoEScale=Base, Params=83M, FLOPs=43G2026.03 | 86.75 | 0.12 | |
| Transformer+attention gate (AG)Scale=Base, Params=20M, FLOPs=47G2026.03 | 86.72 | 0.09 | |
| Transformer+local attention (LA)Scale=Base, Params=18M, FLOPs=40G2026.03 | 86.71 | 0.08 | |
| Transformer+QKNorm (QKN)Scale=Base, Params=18M, FLOPs=43G2026.03 | 86.7 | 0.07 | |
| TransformerScale=Base, Params=18M, FLOPs=43G2026.03 | 86.63 | — | |
| SORT(Transformer+ST+LA+QP+AG+QKN+MoE)Scale=Small, Params=14M, FLOPs=4G2026.03 | 86.56 | 0.44 | |
| HSTUScale=Base, Params=18M, FLOPs=43G2026.03 | 86.44 | -0.19 | |
| TransformerScale=Small, Params=3M, FLOPs=8G2026.03 | 86.12 | — |