verymadmatt

@hsfzxjy thanks for your quick response. the following is the training log. the training resumed from epoch 460. 2020-01-23 08:38:18,729 Namespace(cfg='experiments/cityscapes/seg_hrnet_ocr_w48_train_512x1024_sgd_lr1e-2_wd5e-4_bs_12_epoch484.yaml', local_rank=3, opts=[], seed=304) 2020-01-23 08:38:18,729 Namespace(cfg='experiments/cityscapes/seg_hrnet_ocr_w48_train_512x1024_sgd_lr1e-2_wd5e-4_bs_12_epoch484.yaml', local_rank=2, opts=[], seed=304)...

@PkuRainBow Thanks for your reply. I'm using python 3.6, pytorch 1.1 and 4 P100 GPUs. Others just followed the 'requirements.txt' file. The mIoU increased to 0.8064 when i turned class_balance...

@hsfzxjy @PkuRainBow FYI. I have reproduced the performance of HRNetV2 with the provided config file. The mIoU reached 0.8086 as reported (0.809). I tried HRNetV2+OCR with random_brightness turned on as...

> > @PkuRainBow Thanks for your reply. I'm using python 3.6, pytorch 1.1 and 4 P100 GPUs. Others just followed the 'requirements.txt' file. The mIoU increased to 0.8064 when i...

> @verymadmatt We recommend you to try our "HRNet + OCR" on the other two datasets including PASCAL-Context and LIP. Thanks for your response. So there're no bugs in the...