mAP@0.5: 34.6
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@0.5: 30.1
mAP@0.5: 27.0
D2-Net: Weakly-Supervised Action Localization via Discriminative Embeddings and Denoised Activations
mAP: 3.35mAP IOU@0.5: 6.75mAP IOU@0.75: 3.02mAP IOU@0.95: 0.82
mAP@0.5: 45.3
mAP@0.5: 40.0
mAP@0.5: 40.5avg-mAP (0.1-0.5): 58.2
mAP@0.5: 33.7
mAP@0.5: 26.6
mAP@0.5: 32.4
mAP@0.5: 44.2mAP@0.5:0.95: 27.6
mAP@0.5: 32.2
mAP@0.5:0.95: 21.2
mAP@0.5: 44.5mAP@0.5:0.95: 28.8
mAP@0.5: 45.1mAP@0.5:0.95: 28.1
mAP@0.5: 35.9
mAP@0.5: 43.2mAP@0.5:0.95: 26.8
mAP@0.5: 41.8mAP@0.5:0.95: 26.3
mAP: 3.45mAP IOU@0.5: 6.18mAP IOU@0.75: 3.15mAP IOU@0.95: 0.83
mAP: 3.30mAP IOU@0.5: 2.68mAP IOU@0.75: 0.81mAP IOU@0.95: 3.30
mAP@0.5: 43.9mAP@0.5:0.95: 27.3
mAP@0.5: 30.5