Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/NAT-M3

NAT-M3

Reported on 35 benchmarks across 4 tasks · 1 paper · 5 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology26 results

Neural Architecture SearchonCIFAR-10 Image Classification
Percentage error· uses extra data· 2020-05-12
1.8
best: 1.6 (NAT-M4)
SOTA
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonImageNet
Top-1 Error Rate· 2020-05-12
20.1
best: 16.1 (DeepMAD-50M)
SOTA
Neural Architecture Transfer arXiv:2005.05859
AutoMLonCIFAR-10 Image Classification
Percentage error· uses extra data· 2020-05-12
1.8
best: 1.6 (NAT-M4)
SOTA
Neural Architecture Transfer arXiv:2005.05859
AutoMLonImageNet
Top-1 Error Rate· 2020-05-12
20.1
best: 16.1 (DeepMAD-50M)
SOTA
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonCINIC-10
Accuracy (%)· 2020-05-12
94.3
best: 94.8 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonDTD
Accuracy (%)· 2020-05-12
78.4
best: 79.1 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonFGVC Aircraft
Accuracy (%)· 2020-05-12
90.1
best: 90.8 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonCIFAR-100
Percentage Error· uses extra data· 2020-05-12
12.3
best: 11.7 (DNA-c)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonCIFAR-10
Search Time (GPU days)· uses extra data· 2020-05-12
1
best: 224 (AlphaX-1 (cutout NASNet))
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonFood-101
Accuracy (%)· 2020-05-12
89
best: 89.4 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonOxford 102 Flowers
Accuracy (%)· 2020-05-12
98.1
best: 98.3 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonSTL-10
Accuracy (%)· 2020-05-12
97.8
best: 97.9 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonOxford-IIIT Pet Dataset
Accuracy (%)· 2020-05-12
94.1
best: 94.3 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonStanford Cars
Accuracy (%)· 2020-05-12
92.6
best: 92.9 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
Neural Architecture SearchonImageNet
Accuracy· 2020-05-12
79.9
best: 83.9 (DeepMAD-50M)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonCINIC-10
Accuracy (%)· 2020-05-12
94.3
best: 94.8 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonDTD
Accuracy (%)· 2020-05-12
78.4
best: 79.1 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonFGVC Aircraft
Accuracy (%)· 2020-05-12
90.1
best: 90.8 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonCIFAR-100
Percentage Error· uses extra data· 2020-05-12
12.3
best: 11.7 (DNA-c)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonCIFAR-10
Search Time (GPU days)· uses extra data· 2020-05-12
1
best: 224 (AlphaX-1 (cutout NASNet))
Neural Architecture Transfer arXiv:2005.05859
AutoMLonFood-101
Accuracy (%)· 2020-05-12
89
best: 89.4 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonOxford 102 Flowers
Accuracy (%)· 2020-05-12
98.1
best: 98.3 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonSTL-10
Accuracy (%)· 2020-05-12
97.8
best: 97.9 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonOxford-IIIT Pet Dataset
Accuracy (%)· 2020-05-12
94.1
best: 94.3 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonStanford Cars
Accuracy (%)· 2020-05-12
92.6
best: 92.9 (NAT-M4)
Neural Architecture Transfer arXiv:2005.05859
AutoMLonImageNet
Accuracy· 2020-05-12
79.9
best: 83.9 (DeepMAD-50M)
Neural Architecture Transfer arXiv:2005.05859

Computer Vision9 results

Image ClassificationonCINIC-10
Accuracy· 2020-05-12
94.3
best: 95.8 (VIT-L/16 (Spinal FC, Background))
SOTA
Neural Architecture Transfer arXiv:2005.05859
Image ClassificationonCIFAR-10
Percentage correct· uses extra data· 2020-05-12
98.2
best: 99.5 (ViT-H/14)
Neural Architecture Transfer arXiv:2005.05859
Image ClassificationonCIFAR-10
Top-1 Accuracy· uses extra data· 2020-05-12
98.2
best: 99.12 (Astroformer)
Neural Architecture Transfer arXiv:2005.05859
Image ClassificationonCIFAR-100
Percentage correct· uses extra data· 2020-05-12
87.7
best: 96.08 (EffNet-L2 (SAM))
Neural Architecture Transfer arXiv:2005.05859
Image ClassificationonSTL-10
Percentage correct· uses extra data· 2020-05-12
97.8
best: 99.64 (µ2Net+ (ViT-L/16))
Neural Architecture Transfer arXiv:2005.05859
Image ClassificationonOxford-IIIT Pets
Accuracy· 2020-05-12
94.1
best: 97.1 (EffNet-L2 (SAM))
Neural Architecture Transfer arXiv:2005.05859
Image ClassificationonFood-101
Accuracy· 2020-05-12
89
best: 98.6 (CAP)
Neural Architecture Transfer arXiv:2005.05859
Fine-Grained Image ClassificationonOxford-IIIT Pets
Accuracy· 2020-05-12
94.1
best: 97.1 (EffNet-L2 (SAM))
Neural Architecture Transfer arXiv:2005.05859
Fine-Grained Image ClassificationonFood-101
Accuracy· 2020-05-12
89
best: 98.6 (CAP)
Neural Architecture Transfer arXiv:2005.05859