Not able to reproduce - segmentation result

[yongmayer]

Hello,

Thank you very much for sharing the great work! I was trying to reproduce the segmentation result using bash tools/dist_train.sh configs/ade/ddp_swin_s_2x8_512x512_160k_ade20k.py 1

Unfortunately, I cannot reproduce a similar result (not even close): mIoU after 16k steps is only 21.04 as compared to 37.93 in your log. I cannot figure it out why. Environment more or less similar except for CUDA, for which I don't except that huge difference. Log is attached. 20240424_065202.log

I would appreciate any help! Thank you!

[bqm1111]

Do you find any reason for this? I encounter the same problem.