Neural Architecture Search on ImageNet16-120 NAS-Bench-201 (val)
46.77Accuracyoptimal
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| optimal2021.06 | 46.77 | — | — | — | |
| optimal2021.11 | 46.77 | — | — | — | |
| optimal2022.03 | 46.77 | — | — | — | |
| OptimalSearch Space=NAS-Bench-2012022.04 | 46.77 | — | — | — | |
| Optimum2022.03 | 46.77 | — | — | — | |
| Optimum2024.03 | 46.77 | — | — | — | |
| AG-NetQueries=400, Search Method=Generative LSO2022.03 | 46.73 | — | — | — | |
| Optimal2025.09 | 46.73 | — | — | — | |
| DiNASQueries/Gen.=1922024.03 | 46.66 | — | — | — | |
| BANANASQueries=192, Search Method=Bayesian Optimization2022.03 | 46.65 | — | — | — | |
| BANANASQueries/Gen.=1922024.03 | 46.65 | — | — | — | |
| AG-NetQueries=192, Search Method=Generative LSO2022.03 | 46.64 | — | — | — | |
| AG-NetQueries/Gen.=1922024.03 | 46.64 | — | — | — | |
| CoLLM-NASParadigm=LLM-based NAS methods2025.09 | 46.62 | — | — | — | |
| RZ-NASParadigm=LLM-based NAS methods2025.09 | 46.53 | — | — | — | |
| CR-LSO2022.11 | 46.51 | — | — | — | |
| XGBQueries=192, Search Method=Generative LSO2022.03 | 46.48 | — | — | — | |
| LM-SearcherParadigm=LLM-based NAS methods2025.09 | 46.48 | — | — | — | |
| BO+Queries=192, Search Method=Bayesian Optimization2022.03 | 46.43 | — | — | — | |
| Bayesian Opt.Queries/Gen.=1922024.03 | 46.43 | — | — | — | |
| AG-NetQueries=100, Search Method=Generative LSO, topk=12022.03 | 46.42 | — | — | — | |
| XGB + RankingQueries=192, Search Method=Generative LSO2022.03 | 46.4 | — | — | — | |
| DrNAS2021.11 | 46.37 | — | — | — | |
| β-DARTSCost (hours)=3.22022.03 | 46.37 | — | — | — | |
| B-DARTS2024.03 | 46.37 | — | — | — | |
| Arch2vec + BOQueries=100, Search Method=Bayesian Optimization2022.03 | 46.34 | — | — | — | |
| arch2vec-BO2022.11 | 46.34 | — | — | — | |
| AG-NetQueries=96, Search Method=Generative LSO2022.03 | 46.31 | — | — | — | |
| CARL2026.05 | 46.27 | — | — | — | |
| ConvNP2026.05 | 46.24 | — | — | — | |
| NP2026.05 | 46.19 | — | — | — | |
| Evolutionary AlgorithmParadigm=conventional search algorithms2025.09 | 46.19 | — | — | — | |
| RF2026.05 | 46.15 | — | — | — | |
| BaLeNAS-TFOptimization=Training-free proxies, Proxy=Synflow2021.11 | 46.14 | — | — | — | |
| CATCH-metaAlgorithm Type=Meta-learning based NAS2020.07 | 46.07 | 0.6 | 5 | 22.5 | |
| RE2026.05 | 46.01 | — | — | — | |
| CAP2026.05 | 46 | — | — | — | |
| Random SearchQueries/Gen.=1922024.03 | 45.97 | — | — | — | |
| RS+Queries=192, Search Method=Random2022.03 | 45.87 | — | — | — | |
| ReNAS2022.11 | 45.85 | — | — | — | |
| RR2026.05 | 45.78 | — | — | — | |
| XGB2026.05 | 45.73 | — | — | — | |
| Random SearchParadigm=conventional search algorithms2025.09 | 45.52 | — | — | — | |
| BOHB2022.11 | 45.44 | — | — | — | |
| MLP2026.05 | 45.42 | — | — | — | |
| RS2026.05 | 45.4 | — | — | — | |
| BaLeNASOrder=2nd-order2021.11 | 45.39 | — | — | — | |
| GENIUSParadigm=LLM-based NAS methods2025.09 | 45.29 | — | — | — | |
| Reinforcement LearningParadigm=conventional search algorithms2025.09 | 45.22 | — | — | — | |
| BaLeNASOrder=1st-order2021.11 | 45.19 | — | — | — | |
| PRE-NASSearch Space=NAS-Bench-2012022.04 | 45.16 | — | — | — | |
| AmoebaNetSearch Space=NAS-Bench-2012022.04 | 45.15 | — | — | — | |
| LR2026.05 | 45.07 | — | — | — | |
| REINFORCESearch Space=NAS-Bench-2012022.04 | 45.05 | — | — | — | |
| REINFORCE2022.11 | 45.05 | — | — | — | |
| LLMaticParadigm=LLM-based NAS methods2025.09 | 44.98 | — | — | — | |
| ENASAlgorithm Type=One-shot2020.07 | 44.92 | 0.51 | 5.18 | 13.66 | |
| EG-NASParadigm=one-shot NAS2025.09 | 44.89 | — | — | — | |
| DARTS-Cost (hours)=3.22022.03 | 44.87 | — | — | — | |
| DARTS-Paradigm=one-shot NAS2025.09 | 44.87 | — | — | — | |
| GDASAlgorithm Type=One-shot2020.07 | 44.81 | 0.97 | 17 | 29.23 | |
| SGNASSearch Method=Supernet2022.03 | 44.65 | — | — | — | |
| SGNAS2024.03 | 44.65 | — | — | — | |
| Random SearchSearch Space=NAS-Bench-2012022.04 | 44.45 | — | — | — | |
| RS2022.11 | 44.45 | — | — | — | |
| BOHBSearch Space=NAS-Bench-2012022.04 | 44.26 | — | — | — | |
| R-NASAlgorithm Type=One-shot2020.07 | 44.12 | 1.04 | 5.94 | 10.39 | |
| Zero-cost NAS2021.11 | 43.24 | — | — | — | |
| SNAS2022.03 | 42.84 | — | — | — | |
| FairNAS2022.11 | 41.9 | — | — | — | |
| FairNASParadigm=one-shot NAS2025.09 | 41.9 | — | — | — | |
| GDAS2021.11 | 41.7 | — | — | — | |
| GDASSearch Space=NAS-Bench-2012022.04 | 41.7 | — | — | — | |
| GDASCost (hours)=8.72022.03 | 41.59 | — | — | — | |
| GDAS2022.11 | 41.59 | — | — | — | |
| GDAS2021.06 | 41.28 | — | — | — | |
| SETNAlgorithm Type=One-shot2020.07 | 41.04 | 1.64 | 20.33 | 35.69 | |
| PC-DARTS2022.03 | 40.83 | — | — | — | |
| PC-DARTSParadigm=one-shot NAS2025.09 | 40.83 | — | — | — | |
| DSNAS2022.03 | 40.61 | — | — | — | |
| iDARTS2021.06 | 40.38 | — | — | — | |
| iDARTS2022.03 | 40.38 | — | — | — | |
| Random baseline2021.11 | 33.34 | — | — | — | |
| SETN2021.06 | 33.06 | — | — | — | |
| SETN2021.11 | 33.06 | — | — | — | |
| SETNSearch Space=NAS-Bench-2012022.04 | 32.54 | — | — | — | |
| RandomNAS2021.11 | 31.63 | — | — | — | |
| RSPSSearch Space=NAS-Bench-2012022.04 | 31.56 | — | — | — | |
| RandomNAS2021.06 | 27.22 | — | — | — | |
| DARTS-V2Algorithm Type=One-shot2020.07 | 26.29 | 0 | 22.14 | 39.05 | |
| DARTS-V1Algorithm Type=One-shot2020.07 | 23.66 | 0 | 4.55 | 9.45 | |
| DARTS2021.06 | 16.43 | — | — | — | |
| DARTSOrder=1st-order2021.11 | 16.43 | — | — | — | |
| DARTSOrder=2nd-order2021.11 | 16.43 | — | — | — | |
| DARTS(1st)Cost (hours)=3.22022.03 | 16.43 | — | — | — | |
| DARTS(2nd)Cost (hours)=10.22022.03 | 16.43 | — | — | — | |
| DARTS-V1Search Space=NAS-Bench-2012022.04 | 16.43 | — | — | — | |
| DARTS-V2Search Space=NAS-Bench-2012022.04 | 16.43 | — | — | — | |
| ENASSearch Space=NAS-Bench-2012022.04 | 16.43 | — | — | — | |
| DARTS2022.11 | 16.43 | — | — | — |