xwd-szu

Nice works Can you provide the code of T-SNE visualization.I really need it. Thank you for your to rely!

The problem of image shape

4

when use "python3 train.py --config configs/voc_cac_deeplabv3+_resnet50_1over8_datalist0.json" with two GPUS to run , i find that one GPUS imput image size is (1,3,335,500) while another is (1,3,366,500) . in this case...

Reproduction issues

5

Nice Work！！ Thank you very much for your work! However, when I reproduce your code, I never reach the metric of your paper, is there any training trick ? Looking...

Problems of PatchNCELoss

4

Hi, thanks for your excellent work. In PatchNCELoss , the dim of input feat_q and feat_k is what and each dim means?

Gpus

4

i use three gpus to train，but get an error，cound you give me some segguesttion

nice work！ Why the original image of H*W*3，What is the purpose of this？

About 4D Cost Volume Visualization

1

Nice Works ! I tried to visualize 4D cost volume by myself, but I can't achieve the appearance in your paper, can you provide your visualization code? Looking forward to...

about non-differentiable top-k

3

Nice Works! Being very interested in your work，But I found that your top-k ranking removes the differentiable top-k method and uses softmax instead, what is the reason for this ?

wil provide this code not based on mmdetection

1

Nice wordks！ could you provide this code not based on mmdetection ? Or could you provide a detailed README

When I use the pre-trained weights（SRCC_758_LCC_765.pth） you provided for inference, I get the following results: test, accuracy: 0.8004852849092048, lcc_mean: 0.7643782940100576, srcc_mean: 0.7552650871245324, validate_losses: 0.04839770108433388 Why is the accuracy only 0.8,...