Closed ghost closed 7 years ago
Hi, Dang,
Thank you for pointing out this issue. You are right, I had been making changes to matconvnet when training the models on different datasets. So this incompatibility issues arises. Now I fix it and update in github. Basically I remove the "usingGlobal" field and add new variables.
On Thu, Jun 8, 2017 at 8:45 AM, Dang Kang notifications@github.com wrote:
Hi aimerykong:
Thanks a lot for releasing this interesting code. When I tried to test the cityscape model with the released code, the following error comes out:
Error using dagnn.Layer/load (line 200) No property usingGlobal for a layer of type dagnn.BatchNorm.
However, such an error does not occur for NYUV2 model. I am wondering is there issues such mismatch between the model file and the code?
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/aimerykong/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop/issues/1, or mute the thread https://github.com/notifications/unsubscribe-auth/AGKZJEr4mZm0Z4OI2U4Mdi5enc4i8Aieks5sCBcTgaJpZM4N0Qtk .
Hi, aimerykong:
Thanks for your fast response! I can confirm this issue has been resolved. And thanks again for sharing.
Hi aimerykong:
Thanks a lot for releasing this interesting code. When I tried to test the cityscape model with the released code, the following error comes out:
Error using dagnn.Layer/load (line 200) No property
usingGlobal
for a layer of typedagnn.BatchNorm
.However, such an error does not occur for NYUV2 model. I am wondering is there issues such mismatch between the model file and the code?