You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I train the model HRNet-W48+OCR with the configuration seg_hrnet_ocr_w48_473x473_sgd_lr7e-3_wd5e-4_bs_40_epoch150, but finally only get 52% mIoU on the dataset LIP. How to deal with it? Or are there any train tips not mentioned in README?
The text was updated successfully, but these errors were encountered:
I train the model HRNet-W48+OCR with the configuration seg_hrnet_ocr_w48_473x473_sgd_lr7e-3_wd5e-4_bs_40_epoch150, but finally only get 52% mIoU on the dataset LIP. How to deal with it? Or are there any train tips not mentioned in README?
The text was updated successfully, but these errors were encountered: