Error when i testing the demo_cityscapes and demo_NYUv2

[xjlnuc]

Hi aimerykong: The error are as follows 错误使用 dagnn.Layer/load (line 200) No property bnorm_moment_type_trn for a layer of type dagnn.BatchNorm.

出错 dagnn.DagNN.loadobj (line 27) block.load(struct(s.layers(l).block)) ;

出错 demo_Cityscapes (line 82) netMat = dagnn.DagNN.loadobj(netMat);

错误使用 dagnn.Layer/load (line 200) No property noise_param_idx for a layer of type dagnn.BatchNorm.

出错 dagnn.DagNN.loadobj (line 28) block.load(struct(s.layers(l).block)) ;

出错 demo_NYUv2 (line 40) netMat = dagnn.DagNN.loadobj(netMat);

[aimerykong]

Hi, it seems you use a different matconvnet toolbox? Do you download and use the official version? I add new layers, so please make sure you are using the one I uploaded in github. Thanks!

On Sat, Sep 16, 2017 at 3:00 AM, xjlnuc notifications@github.com wrote:

Hi aimerykong: The error are as follows 错误使用 dagnn.Layer/load (line 200) No property bnorm_moment_type_trn for a layer of type dagnn.BatchNorm.

出错 dagnn.DagNN.loadobj (line 27) block.load(struct(s.layers(l).block)) ;

出错 demo_Cityscapes (line 82) netMat = dagnn.DagNN.loadobj(netMat);

错误使用 dagnn.Layer/load (line 200) No property noise_param_idx for a layer of type dagnn.BatchNorm.

出错 dagnn.DagNN.loadobj (line 28) block.load(struct(s.layers(l).block)) ;

出错 demo_NYUv2 (line 40) netMat = dagnn.DagNN.loadobj(netMat);

— 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/10, or mute the thread https://github.com/notifications/unsubscribe-auth/AGKZJC4-fhP-ncShM6c-nagnao5cWJCrks5si5wngaJpZM4PZxfC .

[xjlnuc]

Hi aimerykong: Thank you very much for the quick reply，you are right,i have use a different matconvnet toolbox,but when i use the official version,I can compile successfully,and when i use your matconvnet,there are many problems when i compiling for gpu,the problem is

F:/personal file2/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop-master/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop-master/matconvnet/matlab/src/bits/impl/bilinearsampler_gpu.cu(247): warning: variable "backward" was declared but never referenced Detected during:instantiation of "vl::ErrorCode forward_backward<type,backwardData,backwardGrid>(vl::Context &, type , type , type , const type , const type , const type , size_t, size_t, size_t, size_t, size_t, size_t, size_t) [with type=float, backwardData=true, backwardGrid=false]"
(352): here instantiation of "vl::ErrorCode vl::impl::bilinearsampler<vl::VLDT_GPU, type>::backward(vl::Context &, type , type , const type , const type , const type *, size_t, size_t, size_t, size_t, size_t, size_t, size_t) [with type=float]"

i don’t konw what happend And,when test the cpu and gpu,the result are as follows I think I have not compiled successfully,because when i am running the demo_cityspace on this basis,there are errors as follows: I have use the win10+matlab2017a+vs2015+cuda8.0+cudnn5.1,Is it because I am using the wrong version?Can you tell me what version you are using?I would be grateful.

[aimerykong]

Hi, can you compile it without testing. besides writing new functions for the new layers, I didn't add them in the testing scripts. How about compiling using CPU only? If that works, compile using GPU without cudnn.

On Sun, Sep 17, 2017 at 9:34 PM, xjlnuc notifications@github.com wrote:

Hi aimerykong: Thank you very much for the quick reply，you are right,i have use a different matconvnet toolbox,but when i use the official version,I can compile successfully,and when i use your matconvnet,there are many problems when i compiling for gpu,the problem is [image: image] https://user-images.githubusercontent.com/30426708/30529145-f7854276-9c6c-11e7-85ec-1b051574537e.png i don’t konw what happend And,when test the cpu and gpu,the result are as follows [image: image] https://user-images.githubusercontent.com/30426708/30529154-0d8c5b5e-9c6d-11e7-8ff7-1d7e2024e0e2.png [image: image] https://user-images.githubusercontent.com/30426708/30529184-56926988-9c6d-11e7-8e84-481e10092ab4.png I think I have not compiled successfully,because when i am running the demo_cityspace on this basis,there are errors as follows: [image: image] https://user-images.githubusercontent.com/30426708/30529194-7dcfed68-9c6d-11e7-82a6-b4713257a6d4.png I have use the win10+matlab2017a+vs2015+cuda8.0+cudnn5.1,Is it because I am using the wrong version?Can you tell me what version you are using?I would be grateful.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/aimerykong/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop/issues/10#issuecomment-330127880, or mute the thread https://github.com/notifications/unsubscribe-auth/AGKZJFq9VSd4pxJf7oE5fuX_i70C2Wk2ks5sjfLBgaJpZM4PZxfC .

[xjlnuc]

Hi,i can successfully compile cpu,and i have compiled GPU without cudnn,but there is still have the same result when i run the demo_cityspaces.I will try win7 to see if it can be successful. The following is my problem in the compile of the GPU, and I think the code could be the cause of the error： 1. d:\recurrent-scene-parsing-with-perspective-understanding-in-the-loop-master\matconvnet\matlab\src\bits/impl/tinythread.h(654): warning: extra ";" ignored

D:/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop-master/matconvnet/matlab/src/vl_imreadjpeg.cu(1329): warning: statement is unreachable

2. D:/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop-master/matconvnet/matlab/src/bits/impl/bilinearsampler_gpu.cu(247): warning: variable "backward" was declared but never referenced

3. D:/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop-master/matconvnet/matlab/src/bits/impl/bilinearsampler_gpu.cu(247): warning: variable "backward" was declared but never referenced detected during: instantiation of "vl::ErrorCode forward_backward<type,backwardData,backwardGrid>(vl::Context &, type , type , type , const type , const type , const type , size_t, size_t, size_t, size_t, size_t, size_t, size_t) [with type=float, backwardData=false, backwardGrid=false]"
(325): here instantiation of "vl::ErrorCode vl::impl::bilinearsampler<vl::VLDT_GPU, type>::forward(vl::Context &, type , const type , const type *, size_t, size_t, size_t, size_t, size_t, size_t, size_t) [with type=float]"
(364): here

    Thank you very much!

[aimerykong]

These are not error. These are just warnings. So you have successfully compiled with GPU.

On Tue, Sep 19, 2017 at 4:44 AM, xjlnuc notifications@github.com wrote:

Hi,i can successfully compile cpu,and i have compiled GPU without cudnn,but there is still have the same result when i run the demo_cityspaces.I will try win7 to see if it can be successful. The following is my problem in the compile of the GPU, and I think the code could be the cause of the error： 1. d:\recurrent-scene-parsing-with-perspective-understanding-in-the-loop- master\matconvnet\matlab\src\bits/impl/tinythread.h(654): warning: extra ";" ignored

D:/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop- master/matconvnet/matlab/src/vl_imreadjpeg.cu(1329): warning: statement is unreachable

1.

D:/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop- master/matconvnet/matlab/src/bits/impl/bilinearsampler_gpu.cu(247): warning: variable "backward" was declared but never referenced

1.

D:/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop- master/matconvnet/matlab/src/bits/impl/bilinearsampler_gpu.cu(247): warning: variable "backward" was declared but never referenced detected during: instantiation of "vl::ErrorCode forward_backward<type, backwardData,backwardGrid>(vl::Context &, type , type , type , const type , const type , const type , size_t, size_t, size_t, size_t, size_t, size_t, size_t) [with type=float, backwardData=false, backwardGrid=false]" (325): here instantiation of "vl::ErrorCode vl::impl::bilinearsampler<vl::VLDT_GPU, type>::forward(vl::Context &, type , const type , const type *, size_t, size_t, size_t, size_t, size_t, size_t, size_t) [with type=float]" (364): here

Thank you very much!

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/aimerykong/Recurrent-Scene-Parsing-with-Perspective-Understanding-in-the-loop/issues/10#issuecomment-330513648, or mute the thread https://github.com/notifications/unsubscribe-auth/AGKZJL3cmzUO1HRpFT1YFxA2r0lyQ4Fhks5sj6kfgaJpZM4PZxfC .