Neural Architecture Search on NASBench-101 (val)
95.06AccuracyOptimum
Evaluation results
| Method | Links | |
|---|---|---|
| Optimum2026.02 | 95.06 | |
| LayerNAS2023.04 | 95.05 | |
| DiNASNumber of queries=150–192, Number of runs=102026.02 | 94.98 | |
| REpopulation_size=50, tournament_size=10, objective function=MNAS2023.04 | 94.97 | |
| AG-NetNumber of queries=150–192, Number of runs=102026.02 | 94.9 | |
| RS2023.04 | 94.8 | |
| PPOtrain_batch_size=16, update_batch_size=8, num_updates_per_feedback=10, objective function=MNAS2023.04 | 94.76 | |
| BANANASNumber of queries=150–192, Number of runs=102026.02 | 94.73 | |
| Bayesian Opt.Number of queries=150–192, Number of runs=102026.02 | 94.57 | |
| DGPONumber of queries=~2K, Number of seeds=32026.02 | 94.5 | |
| Reg. Evol.Number of queries=150–192, Number of runs=102026.02 | 94.47 | |
| Optimal2023.04 | 94.32 | |
| Random SearchNumber of queries=150–192, Number of runs=102026.02 | 94.31 |