Weihao

icon indicating a website link https://scholar.google.com/citations?user=d7i9KuIAAAAJ&hl=zh-CN

icon company Shanghai Jiao Tong University, China icon location 800 Dongchuan RD., Minhang Dist., Shanghai, China icon location I am Weihao Yan, a PhD student in Control Science and Engineering at Shanghai Jiao Tong University.

Results 5 issues of Weihao

About the coefficient 1000.0 in DKD.py

2 comment

in line 21 - 26 in DKD.py > pred_teacher_part2 = F.softmax( > logits_teacher / temperature - 1000.0 * gt_mask, dim=1 > ) > log_pred_student_part2 = F.log_softmax( > logits_student / temperature...

关于optimizer的问题

7 comment

我发现STDC中的optimizer与BiSeNet的optimizer相比多传入了boundary_loss_func,输出看了下，请问是因为在训练过程中对初始设置的融合权重0.6,0.3,0.1进行优化吗？

How to split the datasets (cityscapes) into labeled and unlabeled?

2 comment

Hi, nice work and I am going to reproduce the experimental results. I want to know how you split the dataset into the labeld and unlabeled parts? Is it random...

Could you please the model, config, code on synthia-to-cityscapes?

Hi, nice work! And I am going to reimplement your work and make comparison with it in my paper. But I found the Deeplab model is different in Adaptseg and...

Request for Coarse-Grained Geographic Distribution Information of the Mapillary Vistas Dataset

Hello and thank you for maintaining this dataset. I am currently conducting research on cross-region robustness and domain generalization in semantic segmentation. The Mapillary Vistas Dataset is particularly valuable for...