Predicting Transformer Generalization on MNIST-Transformers (80% threshold)
0.892Kendall's TauTransformer-NFN Quasi
Evaluation results
| Method | Links | |
|---|---|---|
| Transformer-NFN QuasiExtra Parameters (%)=4.54% params ++, Number of runs=52026.04 | 0.892 | |
| Transformer-NFN largeExtra Parameters (%)=57.66% params ++, Number of runs=52026.04 | 0.89 | |
| Transformer-NFNNumber of runs=52026.04 | 0.888 | |
| XGBoostNumber of runs=52026.04 | 0.884 | |
| MLPNumber of runs=52026.04 | 0.873 | |
| STATNNNumber of runs=52026.04 | 0.856 | |
| LightGBMNumber of runs=52026.04 | 0.83 | |
| Random ForestNumber of runs=52026.04 | 0.759 |