xmojiao / deeplab_v2

基于v2版本的deeplab,使用VGG16模型,在VOC2012,Pascal-context,NYU-v2等多个数据集上进行训练
162 stars 103 forks source link

seg_accuracy_layer.cpp:87] Unexpected label 38. num: 0. row: 4. col: 6 #5

Open yulijun1234 opened 6 years ago

yulijun1234 commented 6 years ago

作者你好!我用自己的数据集按照您给出的方法训练deeplab时,出现了这个错误,请问能给出指导意见吗?

xmojiao commented 6 years ago

@yulijun1234 请问解决问题了吗,这应该是你的label制作的问题

yanzhou-li commented 3 years ago

作者你好!我用自己的数据集按照您给出的方法训练deeplab时,出现了这个错误,请问能给出指导意见吗?

@yulijun1234, @xmojiao, 请问你解决了吗,制作的label一定要从像素1开始吗?如果是两个类别,前景是10(pixels)的话,那该如何调整标签或者代码,谢谢!

xmojiao commented 3 years ago

不好意思呐,已经不做这个方向了,可能帮不到你

---Original--- From: @.> Date: Mon, Nov 1, 2021 15:31 PM To: @.>; Cc: "Jiao @.**@.>; Subject: Re: [xmojiao/deeplab_v2] seg_accuracy_layer.cpp:87] Unexpected label 38. num: 0. row: 4. col: 6 (#5)

作者你好!我用自己的数据集按照您给出的方法训练deeplab时,出现了这个错误,请问能给出指导意见吗?

@yulijun1234, @xmojiao, 请问你解决了吗,制作的label一定要从像素1开始吗?如果是两个类别,前景是10(pixels)的话,那该如何调整标签或者代码,谢谢!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.