A Continual Development Methodology for Large-scale Multitask Dynamic ML Systems

Open paper
Benchmark Model Rank Results
domain-generalization-on-imagenet-aµ2Net+ (ViT-L/16)#3Top-1 accuracy %: 84.53
fine-grained-image-classification-on-caltechµ2Net+ (ViT-L/16)#4Top-1 Error Rate: 4.06%
fine-grained-image-classification-on-food-101µ2Net+ (ViT-L/16)#8Accuracy: 91.47
fine-grained-image-classification-on-oxford-2µ2Net+ (ViT-L/16)#3Accuracy: 95.5
fine-grained-image-classification-on-stanford-1µ2Net+ (ViT-L/16)#4Accuracy: 93.5%
image-classification-on-dtdµ2Net+ (ViT-L/16)#3Accuracy: 82.23
image-classification-on-emnist-lettersµ2Net+ (ViT-L/16)#7Accuracy: 95.03
image-classification-on-eurosatµ2Net+ (ViT-L/16)#3Accuracy (%): 99.22
image-classification-on-inaturalist-2018µ2Net+ (ViT-L/16)#12Top-1 Accuracy: 80.97
image-classification-on-malaria-datasetµ2Net+ (ViT-L/16)#2Acc. (test): 97.46%
image-classification-on-places365µ2Net+ (ViT-L/16)#4Top 1 Accuracy: 59.15
image-classification-on-stl-10µ2Net+ (ViT-L/16)#1Percentage correct: 99.64
long-tail-learning-on-imagenet-ltµ2Net+ (ViT-L/16)#2Top-1 Accuracy: 82.5
scene-classification-on-uc-merced-land-useµ2Net+ (ViT-L/16)#1Accuracy (%): 100

Rank counts only results with a code link.