search

yyliu01 / PS-MT

[CVPR'22] Perturbed and Strict Mean Teachers for Semi-supervised Semantic Segmentation
https://arxiv.org/pdf/2111.12903.pdf
MIT License
187 stars 17 forks source link

about The results on classic val set based on resnet 50 as backbone? #36

Closed yang654123 closed 1 year ago

yang654123 commented 1 year ago

hi ,thank you for your nice work! i have a question about the results on classic val set based on resnet 50 as backbone,so do you have it? thanks a lot.

yyliu01 commented 1 year ago

Hi @yang654123,

The result of ResNet50 has been reported in Table 1 & 2 of the main paper. Is that what you want?

Cheers, Yuyuan

yang654123 commented 1 year ago

Hi @yang654123,

The result of ResNet50 has been reported in Table 1 & 2 of the main paper. Is that what you want?

Cheers, Yuyuan

sorry，i want the results of classic val set based on resnet50 not blender val set.

yyliu01 commented 1 year ago

Hi @yang654123

I'm confused, I think the validation dataset should be unique for all set-up. In case you are talking about the training set, the ResNet50 results are reported in Table 3.

Could you please provide more details, like which partition protocol or which dataset you are talking about.

Cheers, Yuyuan

yang654123 commented 1 year ago

sorry, i mean the results of 1464labels origin high resolution dataset which also named of classic dataset ,also called official dataset.

yang654123 commented 1 year ago

Hi @yang654123

I'm confused, I think the validation dataset should be unique for all set-up. In case you are talking about the training set, the ResNet50 results are reported in Table 3.

Could you please provide more details, like which partition protocol or which dataset you are talking about.

Cheers, Yuyuan haha I see,but do you have other results on particial protocol like 92,183,366,732 and so on .

yyliu01 commented 1 year ago

Hi @yang654123

The 1464 labels results under ResNet50 are in Tab 3, while we only have resnet101 results for the splits you mentioned, as we followed CPS table. Sorry for the inconvenience.

Cheers, Yuyuan