metadata
license: openrail
Pretrained models of our method MultiAugs
Title: Boosting Semi-Supervised 2D Human Pose Estimation by Revisiting Data Augmentation and Consistency Training
Paper link: waiting to come ...
Code link: https://github.com/hnuzhy/MultiAugs
COCO1K / COCO5K / COCO10K
trained on partly labeled (1k, 5k or 10k) COCO train-set and left unlabeled COCO train-set
- ResNet-18 (256x192, COCO1K, 30 epochs)
- ResNet-18 (256x192, COCO5K, 70 epochs)
- ResNet-18 (256x192, COCO10K, 100 epochs)
COCOall + COCOunlabel
trained on labeled COCO train-set and unlabeled COCO unlabeled-set
- ResNet-50 (256x192, 400 epochs)
- ResNet-101 (256x192, 400 epochs)
- HRNet-W48 (384x288, 400 epochs)
MPII + AIC
trained on labeled MPII train-set and unlabeled AIC train-set
- HRNet-W32 (256x256, 400 epochs)