mAP@0.5: 30.1mAP@0.1:0.7: 37.8mAP@0.1:0.5: 46.9
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: 32.2mAP@0.1:0.7: 40.9mAP@0.1:0.5: 50.3
mAP@0.5: 26.6
avg-mAP (0.1-0.5): 53.2avg-mAP (0.3-0.7): 33.4avg-mAP (0.1:0.7): 42.6
avg-mAP (0.1-0.5): 52.1avg-mAP (0.3-0.7): 32.4avg-mAP (0.1:0.7): 41.5
avg-mAP (0.1-0.5): 52.0avg-mAP (0.3-0.7): 32.4avg-mAP (0.1:0.7): 42.3
mAP@0.5: 21.2mAP@0.1:0.7: -
mAP@0.5: 27mAP@0.1:0.7: 35.3mAP@0.1:0.5: 43.6
mAP@0.5: 23.1mAP@0.1:0.7: 32.4
D2-Net: Weakly-Supervised Action Localization via Discriminative Embeddings and Denoised Activations
mAP@0.5: 35.9mAP@0.1:0.7: -mAP@0.1:0.5: 51.4
mAP@0.5: 6.8mAP@0.1:0.7: -
avg-mAP (0.1-0.5): 71.6avg-mAP (0.3-0.7): 51.1avg-mAP (0.1:0.7): 60.3
avg-mAP (0.1-0.5): 62.7avg-mAP (0.3-0.7): 44.5avg-mAP (0.1:0.7): 52.8
mAP@0.5: 20.3mAP@0.1:0.7: 31.6
avg-mAP (0.1-0.5): 55.6avg-mAP (0.3-0.7): 35.4avg-mAP (0.1:0.7): 44.8
avg-mAP (0.1-0.5): 57.4avg-mAP (0.3-0.7): 38.0avg-mAP (0.1:0.7): 47.0
mAP@0.5: 30.5mAP@0.1:0.7: 41.2mAP@0.1:0.5: 51.2
mAP@0.5: 24.5mAP@0.1:0.7: -
avg-mAP (0.1-0.5): 47.0avg-mAP (0.3-0.7): 28.8avg-mAP (0.1:0.7): 37.8
mAP@0.5: 13.7mAP@0.1:0.7: -
mAP@0.5: 22.8mAP@0.1:0.7: -
mAP@0.5: 16.9mAP@0.1:0.7: 27.0
mAP@0.5: 23.9mAP@0.1:0.7: -
mAP@0.5: 28.8mAP@0.1:0.7: 37.0mAP@0.1:0.5: 45.6
avg-mAP (0.1-0.5): 44.9avg-mAP (0.3-0.7): 30.4avg-mAP (0.1:0.7): 37.7
Weakly-Supervised Temporal Action Detection for Fine-Grained Videos with Hierarchical Atomic Actions
mAP: 4.10mAP IOU@0.5: 7.05mAP IOU@0.75: 3.95mAP IOU@0.95: 1.14
avg-mAP (0.1:0.7): 47.7
avg-mAP (0.1-0.5): 51.6avg-mAP (0.3-0.7): 32.9avg-mAP (0.1:0.7): 41.9