search

Haochen-Wang409 / U2PL

[CVPR'22 & IJCV'24] Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels & Using Unreliable Pseudo-Labels for Label-Efficient Semantic Segmentation
Apache License 2.0
436 stars 61 forks source link

请问 训练中带标签数据必须和无标签数据数量相等吗? #169

Open DongLZY opened 6 months ago

DongLZY commented 6 months ago

你好我看到代码中有这样的判断 assert len(loader_l) == len(loader_u), f"labeled data {len(loader_l)} unlabeled data {len(loader_u)}, imbalance!" 意思是无标签样本必须和有标签样本数量一致吗?

Haochen-Wang409 commented 6 months ago

我们设置了重复采样使得无标签和有标签数据的总iter一致。