mAP: 27.7
Papers with Code · CC BY-SA 4.0
Zero-Shot Learning
Zero-shot learning (ZSL) is a model's ability to detect classes never seen during training. The condition is that the classes are not known during supervised learning.
Earlier work in zero-shot learning use attributes in a two-step approach to infer unknown classes. In the computer vision context, more recent advances learn mappings from image feature space to semantic space. Other approaches learn non-linear multimodal embeddings. In the modern NLP context, language models can be evaluated on downstream tasks without fine tuning.
Benchmark datasets for zero-shot learning include aPY, AwA, and CUB, among others.
( Image credit: Prototypical Networks for Few shot Learning in PyTorch )
Further readings:
논문
mAP IOU@0.5: 50.07mAP: 33.85mAP IOU@0.75: 34.78mAP IOU@0.95: 8.29
mAP: 29.6mAP IOU@0.5: 46.4mAP IOU@0.75: 29.5mAP IOU@0.95: 7.6
mAP IOU@0.5: 50.36mAP: 34.09mAP IOU@0.75: 34.60mAP IOU@0.95: 9.02
mAP: 17.57
mAP: 29.04mAP IOU@0.5: 45.44mAP IOU@0.75: 28.82mAP IOU@0.95: 6.79
mAP: 18.24mAP IOU@0.5: 29.07mAP IOU@0.75: 17.66mAP IOU@0.95: 5.07
Average mAP: 21.76Recall@1x (tIoU=0.5): 42.54
mAP IOU@0.5: 39.12mAP: 32.26
mAP IOU@0.5: 50.56mAP: 34.26mAP IOU@0.75: 34.75mAP IOU@0.95: 9.37
mAP IOU@0.5: 50.94mAP: 34.75mAP IOU@0.75: 35.61mAP IOU@0.95: 7.98
mAP IOU@0.5: 46.45mAP: 30.03mAP IOU@0.75: 29.96mAP IOU@0.95: 8.02
mAP: 23.8mAP IOU@0.5: 37.1mAP IOU@0.75: 23.7mAP IOU@0.95: 5.9
mAP IOU@0.5: 80.9mAP IOU@0.3: 89.7mAP IOU@0.4: 86.7mAP IOU@0.6: 71.0
mAP: 6.75mAP IOU@0.5: 10.65mAP IOU@0.75: 6.43mAP IOU@0.95: 2.50
mAP IOU@0.5: 51.03mAP: 34.24mAP IOU@0.75: 35.17mAP IOU@0.95: 7.44
mAP IOU@0.5: 48.26mAP: 31.11mAP IOU@0.75: 33.16mAP IOU@0.95: 3.27
mAP IOU@0.5: 50.72mAP: 34.48mAP IOU@0.75: 35.28mAP IOU@0.95: 7.87
mAP IOU@0.5: 73.3mAP IOU@0.3: 84.8mAP IOU@0.4: 80.0mAP IOU@0.6: 63.8
mAP IOU@0.5: 72.9mAP IOU@0.3: 83.6mAP IOU@0.4: 80.1mAP IOU@0.6: 62.4
mAP IOU@0.5: 59.3