An Evolutionary Approach to Dynamic Introduction of Tasks in Large-scale Multitask Learning Systems

Open paper
Benchmark Model Rank Results
fine-grained-image-classification-on-sun397µ2Net (ViT-L/16)#1Accuracy: 84.8
image-classification-on-emnist-digitsµ2Net (ViT-L/16)#1Accuracy (%): 99.82
fine-grained-image-classification-on-oxford-102-flowersµ2Net (ViT-L/16)#3Accuracy: 99.61%
image-classification-on-cifar-100µ2Net (ViT-L/16)#3Percentage correct: 94.95
fine-grained-image-classification-on-oxford-iiit-petsµ2Net (ViT-L/16)#4Accuracy: 95.3
image-classification-on-cifar-10µ2Net (ViT-L/16)#4Percentage correct: 99.49
image-classification-on-dtdµ2Net (ViT-L/16)#5Accuracy: 81.0
image-classification-on-eurosatµ2Net (ViT-L/16)#5Accuracy (%): 99.2
fine-grained-image-classification-on-caltech-101µ2Net (ViT-L/16)#8Top-1 Error Rate: 7%
image-classification-on-kuzushiji-mnistµ2Net (ViT-L/16) (extra data)#8Accuracy: 98.68
image-classification-on-mnistµ2Net (ViT-L/16)#48Accuracy: 99.75
image-classification-on-imagenetµ2Net (ViT-L/16)#133Top 1 Accuracy: 86.74%

Rank counts only results with a code link.