Jet Tagging on Quark-Gluon Tagging (test)
0.923AUCParT
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ParTPID Information=full, Pre-training=JETCLASS fine-tuned2022.02 | 0.923 | 0.852 | — | — | — | |
| ParTPID Information=full, Pre-training=from scratch2022.02 | 0.9203 | 0.849 | — | — | — | |
| LorentzNetPID Information=full, Pre-training=from scratch2022.02 | 0.9156 | 0.844 | — | — | — | |
| ParTPID Information=exp, Pre-training=JETCLASS fine-tuned2022.02 | 0.9151 | 0.843 | — | — | — | |
| PCTPID Information=full, Pre-training=from scratch2022.02 | 0.914 | 0.841 | — | — | — | |
| ABCNetPID Information=full, Pre-training=from scratch2022.02 | 0.9126 | 0.84 | — | — | — | |
| ParTPID Information=exp, Pre-training=from scratch2022.02 | 0.9121 | 0.84 | — | — | — | |
| ParticleNetPID Information=exp, Pre-training=from scratch2022.02 | 0.9116 | 0.84 | — | — | — | |
| ParticleNetPID Information=exp, Pre-training=JETCLASS fine-tuned2022.02 | 0.9115 | 0.839 | — | — | — | |
| rPCNPID Information=exp, Pre-training=from scratch2022.02 | 0.9081 | — | — | — | — | |
| PFNPID Information=full, Pre-training=from scratch2022.02 | 0.9052 | — | — | — | — | |
| PFNPID Information=exp, Pre-training=from scratch2022.02 | 0.9005 | — | — | — | — | |
| P-CNNPID Information=exp, Pre-training=from scratch2022.02 | 0.9002 | 0.827 | 34.7 | 91 | — | |
| TransformerLayers=2, # Params=3,529, FLOPs=4,942,1612025.10 | 0.8957 | 0.8209 | — | — | 6.32 | |
| SAL-TLayers=2, # Params=6,171, FLOPs=2,095,7612025.10 | 0.8924 | 0.8177 | — | — | 6.07 | |
| TransformerLayers=1, # Params=1,941, FLOPs=2,479,7612025.10 | 0.8913 | 0.8164 | — | — | 5.98 | |
| LinformerLayers=2, # Params=13,129, FLOPs=1,087,7612025.10 | 0.8906 | 0.816 | — | — | 5.94 | |
| SAL-TLayers=1, # Params=3,196, FLOPs=739,7612025.10 | 0.8888 | 0.8134 | — | — | 5.76 | |
| LinformerLayers=1, # Params=6,741, FLOPs=552,5612025.10 | 0.8882 | 0.8136 | — | — | 5.77 |