Image Classification on ImageNet-1k random 100-class subsets 1.0 (test)
86.56Accuracy (L10 Test)PolarAdamW
Evaluation results
| Method | Links | |
|---|---|---|
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=20242026.05 | 86.56 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=1232026.05 | 86.1 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=mean2026.05 | 85.78 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=72026.05 | 85.42 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=20242026.05 | 85.14 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=422026.05 | 85.03 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=1232026.05 | 84.97 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=mean2026.05 | 84.41 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=72026.05 | 83.82 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=422026.05 | 83.72 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=20242026.05 | 80.75 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=mean2026.05 | 79.98 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=72026.05 | 79.95 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=422026.05 | 79.8 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=DeiT-style recipe, Epochs=300, Seed=1232026.05 | 79.42 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=1232026.05 | 73.75 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=20242026.05 | 73.65 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=mean2026.05 | 73.47 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=72026.05 | 73.37 | |
| PolarAdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=422026.05 | 73.11 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=72026.05 | 72.17 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=1232026.05 | 71.53 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=mean2026.05 | 71.53 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=20242026.05 | 71.31 | |
| MuonBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=422026.05 | 71.13 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=1232026.05 | 64.56 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=20242026.05 | 64.49 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=mean2026.05 | 63.97 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=72026.05 | 63.73 | |
| AdamWBackbone=DeiT-Tiny, Training Recipe=minimal recipe, Epochs=100, Seed=422026.05 | 63.08 |