Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 13:40
    rolarsen commented #133
  • Mar 27 14:27
    smistad updated the wiki
  • Mar 27 14:26
    smistad updated the wiki
  • Mar 27 14:25
    smistad updated the wiki
  • Mar 27 14:12
    smistad updated the wiki
  • Mar 27 14:10
    smistad updated the wiki
  • Mar 27 13:59
    smistad updated the wiki
  • Mar 27 09:23
    smistad updated the wiki
  • Mar 26 18:52
    rolarsen commented on f1e4ab8
  • Mar 26 18:31
    rolarsen commented #133
  • Mar 26 15:48
    rolarsen opened #133
  • Mar 26 15:37

    smistad on master

    added pipeline attributes for C… added more renderers as process… (compare)

  • Mar 26 15:34

    smistad on master

    added ability to set zoom level… (compare)

  • Mar 24 16:10

    smistad on master

    working syntax highlighting on … (compare)

  • Mar 24 15:16

    smistad on master

    do not call update a disabled r… (compare)

  • Mar 24 15:08

    smistad on master

    added ubuntu mono font and fast… fixed bug in getDirectoryList w… improved pipeline system and ed… and 1 more (compare)

  • Mar 18 19:43

    smistad on master

    added deleteMetadata function t… added metadata blacklist to Met… (compare)

  • Mar 18 13:37
    rolarsen closed #132
  • Mar 18 13:00

    smistad on master

    yellow color on warnings on win… removed a cout an added a getWi… fixed initialization issue with… and 1 more (compare)

  • Mar 18 11:19

    smistad on master

    Update README.md (compare)

Erik Smistad
@smistad
I can recommend this tool: http://www.dependencywalker.com/
xiang123
@xiang123
thank for help, Erik!
Erik Smistad
@smistad
Open your dll in this tool and look for red stuff
huangsusan
@huangsusan
Hello ,Eric.When I set the MinimumRadius more than 1 in LiverVesselSegmentation.cpp,the progress seems to get stuck in a loop,and I read your paper,and still do not know why
Erik Smistad
@smistad
The CT-Abdomen image is not made for liver imaging
If I run that example with that image, I get the main vessels in the abdomen and the intenstines
This makes since the intenstines have received contrast agent in this image
You will need an image where contrast has been optimized for the liver and ideally cropped to the liver as wel
wpv1999
@wpv1999
, thanks for pointing it out
, thanks for pointing it out well I talked about tensorflow -FAST
xiang123
@xiang123
IMG_20191115_160707.jpg
looks like I missed a lot of dependency on windows, I am downloading them one by one
Erik Smistad
@smistad
Hm, that shouldn't be necessary @xiang123
Have you been able to run any FAST examples on Windows compiled with msvc 2019?
I starting to wonder if maybe qt 5.9 has issues with msvc 2019
As I said before, I have not tried msvc 2019 yet. I have only used msvc 2017
xiang123
@xiang123
when I try to run testfast, it do run some test cases and failed some
xiang123
@xiang123
I am switching to msvc 2017
I thought your comment on error C2280
did you fix the problem
Severity Code Description Project File Line Suppression State
Error C2280 'tensorflow::FunctionLibraryDefinition &tensorflow::FunctionLibraryDefinition::operator =(const tensorflow::FunctionLibraryDefinition &)': attempting to reference a deleted function tensorflow_CPU E:\FAST\FAST\build\source\FAST\Algorithms\NeuralNetwork\InferenceEngines\external\com_google_absl\absl\meta\type_traits.h 141
I will update my msvc version first to see if problem solved
xiang123
@xiang123
Hi, Erik,have you ever seen this error?Error C3615 constexpr function 'qCountLeadingZeroBits' cannot result in a constant expression qt5
Erik Smistad
@smistad
No, I have not
xiang123
@xiang123
this is m last error in mcvs 2017 "c:\users\lin.xiang\fast_build\build\include\eigen3\unsupported\eigen\cxx11\src/Tensor/TensorMap.h(103): error C2664: 'Eigen::DSizes<int,1>::DSizes(Eigen::DSizes<int,1> &&)': cannot convert argument 1 from 'const Dimensions' to 'const std::array<T,1> &'"
Erik Smistad
@smistad
I have started setting up fast with tensorflow on a new fresh Windows install and will see if I can retrace the steps I used to get this to work on my current machine. I will let you know what I find
Erik Smistad
@smistad
Hi @xiang123 I have now built FAST+ tensorflow cpu on a fresh install on windows 10 with MSCV 2017. It worked fine after fixing two issues which I have uploaded now to smistad/FAST@e185981 The part about the missing protobuf headers which you have already mentioned. And the issue you mentioned before with "Failed to load plugin because.." which was do to a boring mistake of a wrong renaming of the tensorflow DLL.
Erik Smistad
@smistad
screenshot.jpg
xiang123
@xiang123
now it worked on my machine, Thanks Erik @smistad
Erik Smistad
@smistad
Great to hear!
Sorry for not seeing the bugs earlier
Now that for the CPU, the openvino inference engine from Intel is usually faster in my experience, because it can use the integrated gpu on Intel processors
dfpasdel
@dfpasdel
Hi Erik, I have an issue with the installation of FAST with the tensorflow-cuda module. It is not able to find cudnn.h (located in /usr/local/cuda/include) nor the libraries libcudnn.so* (located in /usr/local/cuda/lib64).
The log says: Please specify the comma-separated list of base paths to look for CUDA libraries and headers. [Leave empty to use the default]: Could not find any cudnn.h matching version '7' in any subdirectory:
''
'include'
'include/cuda'
'include/*-linux-gnu'
'extras/CUPTI/include'
'include/cuda/CUPTI'
of:
'/usr/'
Do you have any idea? Thanks
Erik Smistad
@smistad
First of all: have you installed cudnn?
If so by deb package, or by other means?
dfpasdel
@dfpasdel
Yes, I have installed version 7.6.4 with runfile
Erik Smistad
@smistad
Ok, I always install by package
You need to find out where your system has put the cudnn header file
search your system for cudnn.h and cudnn_v7.h
dfpasdel
@dfpasdel
It is in /usr/local/cuda/include
Erik Smistad
@smistad
what is the name of the file cudnn.h or cudnn_v7.h?
dfpasdel
@dfpasdel
Is is cudnn.h
Erik Smistad
@smistad
ok, I see the paths you listed above or not correct, it is not looking in /usr/local/cuda/include
You can copy it to /usr/include then
sudp cp /usr/local/cuda/include/cudnn.h /usr/include/
dfpasdel
@dfpasdel
I can do that, but the same happens for the library libcudnn.so which is located in /usr/local/cuda/lib64, so I was looking for a more generic solution to fetch headers and libraries where they are. But if is only cudnn.h and libcudnn.so* which are required, copying will do the job. Thanks
Erik Smistad
@smistad
Yes, its only those 2. The best fix would be to fix tensorflow so that it looks in the correct folders..