The source only results of Cityscapes

[chinacui]

Hi. The work is excellent great.

I have downloaded it and trained it on my server (4 GPUs, each GPU with 1 batchsize, 100k iters). I only used the GTA5 data to train the seg network (resnet101) and test on the cityscapes dataset (source only). I found the evaluation results is ~24% mIoU, which have a large gap compared to paper (39%). However, if I change the model from eval() to train() state in the evaluation, I got 34% mIoU. May I ask which state you have used in the paper? If it is eval() state, could you help us to find the potential issues?

Thanks.