pointnet2
pointnet2 copied to clipboard
compilation errors in tf_ops folders
I am using Tensorflow 1.2.1( virtual environment) and cuda 8.0. While running the compile scripts in op folders , I get the following errors:
fatal error: tensorflow/core/framework/op.h: No such file or directory #include "tensorflow/core/framework/op.h" tf_sampling.cpp:6:42: fatal error: tensorflow/core/framework/op.h: No such file or directory #include "tensorflow/core/framework/op.h"
Have you any idea how to solve those errors. Thanks in advance.
?
I think u should change all the paths directory in the codes into tensorflow framework folder on ur pc, i remembered that i faced the same problems when trying to run the training and it might leads you to several similar errors which will need unto do the same solution several time
@Omar-Tag Thanks for your reply. Are you using tensorflow with virtual environment? I changed the path of tensorflow in grouping and sampling foldors of 'tf_ops' but still get the same errors !
@imy16 No i am using ubuntu 16
@Omar-Tag Finally I found the solution ! :D I installed two versions of Tensorflow (one with simple pip (1.0) and the other with virtual environment (1.2)) I was confusing the path of the two installed versions of tensorflow.. Just I fixed the path and compiled the tf ops while the virtual environment is activated. Thanks Omar :)
I wonder if you have solved it. I meet it, try change all the paths, but can't work out.
@UCASJingKun Yes, I solved it. I was just confusing the path of Tensorflow's two installed versions. Try to change all the paths directory in the codes into tensorflow framework folder on your pc and compile the tf ops while the virtual environment is activated if you're using Tensorflow with virtual environment.
thanks, I have solved it.
Dear @Omar-Tag @UCASJingKun
I have the same as your errors. But when I change location tensorflow path it is not changed for me.
This my location
And this my new path I want to change in /.bashrc file
Please give me some your idea
many thank
@ccu1tn you dont change the path of tensorflow , what we mean above , is that you should change the paths that is used in the code to the correct paths relevent to each pc , what happened with @Imy16 abd @UCASJingKun is that they were confused with the existence of two tensorflow's version on their environment so they used the wrong directories
Dear Omar-Tag;
So, how to fix those errors. Do you have any example compile cmsis-nn by mbed. Please could you share for me to do it?
Thank you so much
Vào Th 2, 27 thg 5, 2019 vào lúc 16:56 Omar-Tag [email protected] đã viết:
@ccu1tn https://github.com/ccu1tn you dont change the path of tensorflow , what we mean above , is that you should change the paths that is used in the code to the correct paths relevent to each pc , what happened with @Imy16 https://github.com/Imy16 abd @UCASJingKun https://github.com/UCASJingKun is that they were confused with the existence of two tensorflow's version on their environment so they used the wrong directories
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/charlesq34/pointnet2/issues/54?email_source=notifications&email_token=AKKTSNRQHP25EK4THWUPR2DPXOO3ZA5CNFSM4FNJDJ62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWJHC4A#issuecomment-496136560, or mute the thread https://github.com/notifications/unsubscribe-auth/AKKTSNTZH2JM3DBPPQ4VVUDPXOO3ZANCNFSM4FNJDJ6Q .
@ccu1tn Actually, you need to change the tensorflow path and CUDA path in tf_xxx_compile.sh
.
Hi, could you tell me what the commands are to retrieve both the TensorFlow and CUDA paths in my computer? I'm using Ubuntu 18.04 Thanks!
Usually ~/anaconda/envs/env_name/lib/python2.7/site-packages/tensorflow if using Anaconda.
Hi, could you tell me what the commands are to retrieve both the TensorFlow and CUDA paths in my computer? I'm using Ubuntu 18.04 Thanks!
@usherbob, thanks, but I am not using Anaconda. I would just like to know the commands in a (linux) terminal that would let me retrieve both paths easily. I assume such commands wouldn't change if I were using a virtual environment, though the paths themselves might?