search

xuebinqin / U-2-Net

The code for our newly accepted paper in Pattern Recognition 2020: "U^2-Net: Going Deeper with Nested U-Structure for Salient Object Detection."
Apache License 2.0
8.43k stars 1.45k forks source link

the trained u2net_human_seg.pth is based on the U2net #280

Open Testhjf opened 2 years ago

Testhjf commented 2 years ago

@xuebinqin hi,I have a quertion ? the trained u2net_human_seg.pth is based on the current U2net code （include train code）,is there any change before train？

xuebinqin commented 2 years ago

Yes, there is no change.

On Sat, Jan 8, 2022 at 10:06 AM Testhjf @.***> wrote:

@xuebinqin https://github.com/xuebinqin hi,I have a quertion ? the trained u2net_human_seg.pth is based on the current U2net code （include train code）,is there any change before train？

— Reply to this email directly, view it on GitHub https://github.com/xuebinqin/U-2-Net/issues/280#issuecomment-1007893628, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADSGORJPLQMR7GZVE73NU2LUU7H5XANCNFSM5LQHX64Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

You are receiving this because you were mentioned.Message ID: @.***>

-- Xuebin Qin PhD Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage: https://xuebinqin.github.io/

Testhjf commented 2 years ago

@xuebinqin thanks，and when i train my dataset，the loss is keep 0.19 ，when i change learn rate （0.001 to 0.0001），batchsize is 20，it goes down a little（0.19 to 0.18），but keep 0.18 for 6 hours。is there any problems ？maybe it decrease after more hours ？

xuebinqin commented 2 years ago

you need to double check your code for making sure that your data is correctly loaded. The simplest way is to overfitting a one or two sample images.

On Fri, Jan 14, 2022 at 4:05 PM Testhjf @.***> wrote:

@xuebinqin https://github.com/xuebinqin thanks，and when i train my dataset，the loss is keep 0.19 ，when i change learn rate （0.001 to 0.0001），batchsize is 20，it goes down a little（0.19 to 0.18），but keep 0.18 for 6 hours。is there any problems ？maybe it decrease after more hours ？

— Reply to this email directly, view it on GitHub https://github.com/xuebinqin/U-2-Net/issues/280#issuecomment-1013062316, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADSGORLKJZ7U3GSI6OOULQTUWAGQTANCNFSM5LQHX64Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

You are receiving this because you were mentioned.Message ID: @.***>

-- Xuebin Qin PhD Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage: https://xuebinqin.github.io/