AlexNet

Reported on 6 benchmarks across 5 tasks · 3 papers · 3 SOTA

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

Computer Vision4 results

Object Recognitiononshape bias
shape bias· 2018-11-29
42.9
best: 98.7 (Imagen)
SOTA
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness arXiv:1811.12231
Pedestrian DetectiononCaltech
Reasonable Miss Rate· 2015-01-23
23.3
best: 24.8 (LDCF)
SOTA
Taking a Deeper Look at Pedestrians arXiv:1501.05790
Image ClassificationonObjectNet
Top-1 Accuracy· uses extra data
6.78
best: 82.7 (CoCa)
Image ClassificationonObjectNet
Top-5 Accuracy· uses extra data
17.6
best: 82.1 (ViT-H/14)

Robots1 result

Autonomous VehiclesonCaltech
Reasonable Miss Rate· 2015-01-23
23.3
best: 24.8 (LDCF)
SOTA
Taking a Deeper Look at Pedestrians arXiv:1501.05790

Natural Language Processing1 result

Binary ClassificationonInBreast
AUC· 2017-05-23
0.79
best: 0.947 (WCCNet DenseNet-121)
Deep Multi-instance Networks with Sparse Label Assignment for Whole Mammogram Classification arXiv:1705.08550