Generative Recommendation on Instruments
11.38R@10DIGER
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DIGERvariant=Ours2026.01 | 11.38 | 8.44 | — | — | |
| LETTER2026.01 | 11.22 | 8.31 | — | — | |
| ETEGRec2026.01 | 11.06 | 8.1 | — | — | |
| TIGER2026.01 | 10.58 | 7.97 | — | — | |
| HGN2026.01 | 10.48 | 7.74 | — | — | |
| LightGCN2026.01 | 10 | 7.28 | — | — | |
| P5-CID2026.01 | 9.87 | 7.51 | — | — | |
| P5-SID2026.01 | 9.64 | 7.3 | — | — | |
| SASRec2026.01 | 9.47 | 6.9 | — | — | |
| BERT4Rec2026.01 | 8.22 | 6.08 | — | — | |
| MF2026.01 | 7.35 | 4.12 | — | — | |
| Caser2026.01 | 7.1 | 4.09 | — | — | |
| SID-MLPThroughput (Tput)=3,706, Speedup (Spd.)=8.74×2026.05 | 6.2 | 3.32 | 3.96 | 2.59 | |
| SID-MLP++Throughput (Tput)=4,347, Speedup (Spd.)=10.25×2026.05 | 6.12 | 3.27 | 3.95 | 2.57 | |
| TIGERThroughput (Tput)=320, Speedup (Spd.)=0.75×2026.05 | 6.06 | 3.23 | 3.86 | 2.52 | |
| TIGER-kvThroughput (Tput)=424, Speedup (Spd.)=1.00×2026.05 | 6.06 | 3.23 | 3.86 | 2.52 | |
| AtSpeed-SThroughput (Tput)=286, Speedup (Spd.)=0.68×2026.05 | 6.06 | 3.23 | 3.86 | 2.52 | |
| Mamba2Throughput (Tput)=517, Speedup (Spd.)=1.22×2026.05 | 6.05 | 3.26 | 3.88 | 2.56 | |
| GDNThroughput (Tput)=510, Speedup (Spd.)=1.20×2026.05 | 5.98 | 3.21 | 3.85 | 2.53 | |
| EARNThroughput (Tput)=822, Speedup (Spd.)=1.94×2026.05 | 5.96 | 3.19 | 3.83 | 2.5 | |
| BigRec2026.01 | 5.76 | 4.91 | — | — | |
| NEZHAThroughput (Tput)=3,082, Speedup (Spd.)=7.27×2026.05 | 5.67 | 3.08 | 3.71 | 2.45 | |
| AtSpeed-RThroughput (Tput)=94, Speedup (Spd.)=0.22×2026.05 | 3.75 | 2.04 | 2.46 | 1.63 |