Bamboo: Building Mega-Scale Vision Dataset Continually with Human-Machine Synergy

Open paper

Benchmark	Model	Rank	Results
fine-grained-image-classification-on-caltech	Bamboo (ViT-B/16)	#14	Accuracy: 94.8
fine-grained-image-classification-on-oxford-1	Bamboo (ViT-B/16)	#6	Accuracy: 95.1%
fine-grained-image-classification-on-stanford	Bamboo (ViT-B/16)	#42	Accuracy: 93.9%
fine-grained-image-classification-on-sun397	Bamboo (ViT-B/16)	#3	Accuracy: 79.5
image-classification-on-cifar-10	Bamboo (ViT-B/16)	#53	Percentage correct: 98.2
image-classification-on-cifar-100	Bamboo (ViT-B/16)	#25	Percentage correct: 90.2
image-classification-on-dtd	Bamboo (ViT-B/16)	#4	Accuracy: 81.9
image-classification-on-flowers-102	Bamboo (ViT-B/16)	#4	Accuracy: 99.7
image-classification-on-food-101-1	Bamboo (ViTB/16)	#1	Accuracy (%): 92.9
image-classification-on-objectnet	Vit B/16 (Bamboo)	#21	Top-1 Accuracy: 53.9
image-classification-on-objectnet	ResNet-50 (Bamboo)	#40	Top-1 Accuracy: 38.8
image-classification-on-omnibenchmark	Bamboo-R50	#3	Average Top-1 Accuracy: 45.4