Bongard-HOI: Benchmarking Few-Shot Visual Reasoning for Human-Object Interactions

Open paper
Benchmark Model Rank Results
few-shot-image-classification-on-bongard-hoiHuman (Amateur)#1Avg. Accuracy: 91.42
few-shot-image-classification-on-bongard-hoiMeta-Baseline (ImagNet_R50)#4Avg. Accuracy: 55.82
few-shot-image-classification-on-bongard-hoiMeta-Baseline (MoCov2_R50)#5Avg. Accuracy: 54.30
few-shot-image-classification-on-bongard-hoiMeta-Baseline (Scratch_R50)#6Avg. Accuracy: 54.23
few-shot-image-classification-on-bongard-hoiANIL (ImageNet_R50)#7Avg. Accuracy: 49.74

Rank counts only results with a code link.