Neural Architecture Search on CIFAR-100 NAS-Bench-201 (val)
74.49Accuracyoptimal
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| optimal2021.06 | 74.49 | — | — | — | |
| optimal2021.11 | 74.49 | — | — | — | |
| DrNAS2021.11 | 73.49 | — | — | — | |
| β-DARTSCost (hours)=3.22022.03 | 73.49 | — | — | — | |
| optimal2022.03 | 73.49 | — | — | — | |
| Optimum2022.03 | 73.49 | — | — | — | |
| BANANASQueries=192, Search Method=Bayesian Optimization2022.03 | 73.49 | — | — | — | |
| AG-NetQueries=192, Search Method=Generative LSO2022.03 | 73.49 | — | — | — | |
| AG-NetQueries=400, Search Method=Generative LSO2022.03 | 73.49 | — | — | — | |
| Optimum2024.03 | 73.49 | — | — | — | |
| BANANASQueries/Gen.=1922024.03 | 73.49 | — | — | — | |
| AG-NetQueries/Gen.=1922024.03 | 73.49 | — | — | — | |
| B-DARTS2024.03 | 73.49 | — | — | — | |
| DiNASQueries/Gen.=1922024.03 | 73.49 | — | — | — | |
| Optimal2025.09 | 73.49 | — | — | — | |
| CR-LSO2022.11 | 73.44 | — | — | — | |
| CoLLM-NASParadigm=LLM-based NAS methods2025.09 | 73.44 | — | — | — | |
| Arch2vec + BOQueries=100, Search Method=Bayesian Optimization2022.03 | 73.35 | — | — | — | |
| arch2vec-BO2022.11 | 73.35 | — | — | — | |
| RZ-NASParadigm=LLM-based NAS methods2025.09 | 73.35 | — | — | — | |
| BO+Queries=192, Search Method=Bayesian Optimization2022.03 | 73.26 | — | — | — | |
| Bayesian Opt.Queries/Gen.=1922024.03 | 73.26 | — | — | — | |
| AG-NetQueries=96, Search Method=Generative LSO2022.03 | 73.2 | — | — | — | |
| XGB + RankingQueries=192, Search Method=Generative LSO2022.03 | 73.2 | — | — | — | |
| AG-NetQueries=100, Search Method=Generative LSO, topk=12022.03 | 73.14 | — | — | — | |
| XGBQueries=192, Search Method=Generative LSO2022.03 | 73.1 | — | — | — | |
| ConvNP2026.05 | 72.94 | — | — | — | |
| CARL2026.05 | 72.84 | — | — | — | |
| LM-SearcherParadigm=LLM-based NAS methods2025.09 | 72.82 | — | — | — | |
| BaLeNAS-TFOptimization=Training-free proxies, Proxy=Synflow2021.11 | 72.67 | — | — | — | |
| BOHB2022.11 | 72.59 | — | — | — | |
| CATCH-metaAlgorithm Type=Meta-learning based NAS2020.07 | 72.57 | 0.81 | 4 | 22.5 | |
| RR2026.05 | 72.54 | — | — | — | |
| RE2026.05 | 72.35 | — | — | — | |
| Evolutionary AlgorithmParadigm=conventional search algorithms2025.09 | 72.31 | — | — | — | |
| CAP2026.05 | 72.24 | — | — | — | |
| RF2026.05 | 72.15 | — | — | — | |
| RS+Queries=192, Search Method=Random2022.03 | 72.08 | — | — | — | |
| Random SearchQueries/Gen.=1922024.03 | 72.08 | — | — | — | |
| ReNAS2022.11 | 71.96 | — | — | — | |
| MLP2026.05 | 71.78 | — | — | — | |
| Random SearchParadigm=conventional search algorithms2025.09 | 71.62 | — | — | — | |
| REINFORCE2022.11 | 71.61 | — | — | — | |
| BaLeNASOrder=2nd-order2021.11 | 71.53 | — | — | — | |
| LR2026.05 | 71.51 | — | — | — | |
| Reinforcement LearningParadigm=conventional search algorithms2025.09 | 71.51 | — | — | — | |
| NP2026.05 | 71.48 | — | — | — | |
| RS2026.05 | 71.42 | — | — | — | |
| LLMaticParadigm=LLM-based NAS methods2025.09 | 71.41 | — | — | — | |
| DARTS-Cost (hours)=3.22022.03 | 71.36 | — | — | — | |
| DARTS-Paradigm=one-shot NAS2025.09 | 71.36 | — | — | — | |
| GDASCost (hours)=8.72022.03 | 71.34 | — | — | — | |
| GDAS2022.11 | 71.34 | — | — | — | |
| GDAS2021.11 | 71.14 | — | — | — | |
| XGB2026.05 | 71.09 | — | — | — | |
| GENIUSParadigm=LLM-based NAS methods2025.09 | 70.96 | — | — | — | |
| GDAS2021.06 | 70.95 | — | — | — | |
| FairNAS2022.11 | 70.94 | — | — | — | |
| FairNASParadigm=one-shot NAS2025.09 | 70.94 | — | — | — | |
| RS2022.11 | 70.93 | — | — | — | |
| BaLeNASOrder=1st-order2021.11 | 70.88 | — | — | — | |
| EG-NASParadigm=one-shot NAS2025.09 | 70.78 | — | — | — | |
| iDARTS2021.06 | 70.57 | — | — | — | |
| iDARTS2022.03 | 70.57 | — | — | — | |
| Zero-cost NAS2021.11 | 70.55 | — | — | — | |
| R-NASAlgorithm Type=One-shot2020.07 | 70.39 | 1.36 | 2.26 | 10.39 | |
| GDASAlgorithm Type=One-shot2020.07 | 70.33 | 0.85 | 6.23 | 29.23 | |
| SGNASSearch Method=Supernet2022.03 | 70.28 | — | — | — | |
| SGNAS2024.03 | 70.28 | — | — | — | |
| ENASAlgorithm Type=One-shot2020.07 | 69.99 | 1.03 | 4.26 | 13.66 | |
| SNAS2022.03 | 69.69 | — | — | — | |
| DARTS-V1Algorithm Type=One-shot2020.07 | 68.99 | 1.93 | 2.44 | 9.45 | |
| SETNAlgorithm Type=One-shot2020.07 | 68.01 | 0.21 | 7.74 | 35.69 | |
| PC-DARTS2022.03 | 67.12 | — | — | — | |
| PC-DARTSParadigm=one-shot NAS2025.09 | 67.12 | — | — | — | |
| DARTS-V2Algorithm Type=One-shot2020.07 | 65.06 | 2.95 | 7.91 | 39.05 | |
| RandomNAS2021.11 | 60.99 | — | — | — | |
| Random baseline2021.11 | 60.7 | — | — | — | |
| SETN2021.06 | 58.86 | — | — | — | |
| SETN2021.11 | 58.86 | — | — | — | |
| RandomNAS2021.06 | 52.12 | — | — | — | |
| DSNAS2022.03 | 30.87 | — | — | — | |
| DARTS2021.06 | 15.03 | — | — | — | |
| DARTSOrder=1st-order2021.11 | 15.03 | — | — | — | |
| DARTSOrder=2nd-order2021.11 | 15.03 | — | — | — | |
| DARTS(1st)Cost (hours)=3.22022.03 | 15.03 | — | — | — | |
| DARTS(2nd)Cost (hours)=10.22022.03 | 15.03 | — | — | — | |
| DARTS2022.11 | 15.03 | — | — | — | |
| DARTSParadigm=one-shot NAS2025.09 | 15.03 | — | — | — | |
| ENAS2021.06 | 13.37 | — | — | — | |
| ENAS2021.11 | 13.37 | — | — | — | |
| ENAS2022.11 | 13.37 | — | — | — |