search

xuebinqin / BASNet

Code for CVPR 2019 paper. BASNet: Boundary-Aware Salient Object Detection
MIT License
1.35k stars 249 forks source link

a little bug #12

Closed suruoxi closed 4 years ago

suruoxi commented 4 years ago

https://github.com/NathanUA/BASNet/blob/80ede575fd4d16507bb16db3263488e960e68771/model/BASNet.py#L288

Should be "hd6 = self.relu6d_2(self.bn6d_2(self.conv6d_2(hx)))", right?

xuebinqin commented 4 years ago

Thanks a lot for the correction. You're correct. It is fixed now.

On Wed, Sep 4, 2019 at 5:18 AM Peng Zhang notifications@github.com wrote:

https://github.com/NathanUA/BASNet/blob/80ede575fd4d16507bb16db3263488e960e68771/model/BASNet.py#L288

Should be "hd6 = self.relu6d_2(self.bn6d_2(self.conv6d_2(hx)))", right?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/NathanUA/BASNet/issues/12?email_source=notifications&email_token=ADSGORJNI5T3NNQMTYSDDWLQH6KRDA5CNFSM4ITQU6M2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HJHBPVA, or mute the thread https://github.com/notifications/unsubscribe-auth/ADSGORLWEQSVPL63KHIDRZTQH6KRDANCNFSM4ITQU6MQ .

-- Xuebin Qin PhD Candidate Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage:https://webdocs.cs.ualberta.ca/~xuebin/

mahendraBabuHari commented 4 years ago

I had trained my model using the previous version of BASNet.py. Does that mean that, 1) I need to retrain my model from scratch? 2) Just update the py file without having to retrain?

Thanks.

xuebinqin commented 4 years ago

Sorry about the error. You can use your previous model and that will not make that much difference.

If you want to download and use the new version code. You can copy the bn5d_2 parameters of your previous model to the new model's bn6d_2 and fine tune the model.

On Sat, Oct 5, 2019 at 7:42 PM mahendraBabuHari notifications@github.com wrote:

I had trained my model using the previous version of BASNet.py. Does that mean that,

  1. I need to retrain my model from scratch?
  2. Just update the py file without having to retrain?

Thanks.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/NathanUA/BASNet/issues/12?email_source=notifications&email_token=ADSGORJN4ZK3IWW4HDBXRRDQNE7A7A5CNFSM4ITQU6M2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAN7RBA#issuecomment-538704004, or mute the thread https://github.com/notifications/unsubscribe-auth/ADSGORM4MXHSERJE2KG3R6LQNE7A7ANCNFSM4ITQU6MQ .

-- Xuebin Qin PhD Candidate Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage:https://webdocs.cs.ualberta.ca/~xuebin/