CTR Prediction on Taobao Evaluation Dataset (offline)
69.63AUCEST
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ESTModeling Paradigm=Scalable Unified Modeling, Params(M)=103.81, GFLOPs=18.482026.02 | 69.63 | 0.99 | 65.15 | 0.87 | |
| OneTrans-FModeling Paradigm=Scalable Unified Modeling, Params(M)=110.54, GFLOPs=52.792026.02 | 69.55 | 0.87 | 65.07 | 0.75 | |
| OneTrans-DModeling Paradigm=Scalable Unified Modeling, Params(M)=112.06, GFLOPs=16.192026.02 | 69.41 | 0.68 | 65 | 0.64 | |
| HiFormerModeling Paradigm=Scalable Hierarchical Modeling, Params(M)=113.78, GFLOPs=14.172026.02 | 69.28 | 0.49 | 64.93 | 0.53 | |
| OneTransModeling Paradigm=Scalable Unified Modeling, Params(M)=111.98, GFLOPs=15.942026.02 | 69.21 | 0.38 | 64.77 | 0.27 | |
| MTGRModeling Paradigm=Scalable Unified Modeling, Params(M)=118.79, GFLOPs=10.572026.02 | 69.17 | 0.32 | 64.74 | 0.23 | |
| DCNv2Modeling Paradigm=Traditional DNN Model, Params(M)=15.18, GFLOPs=1.492026.02 | 69.16 | 0.31 | 64.82 | 0.35 | |
| RankMixerModeling Paradigm=Scalable Hierarchical Modeling, Params(M)=136.84, GFLOPs=15.462026.02 | 69.14 | 0.28 | 64.82 | 0.35 | |
| AutoIntModeling Paradigm=Traditional DNN Model, Params(M)=5.20, GFLOPs=1.102026.02 | 69.07 | 0.18 | 64.77 | 0.28 | |
| MLPModeling Paradigm=Base Model, Params(M)=83.13, GFLOPs=9.642026.02 | 68.95 | — | 64.59 | — |