jiasenlu / NeuralBabyTalk

Pytorch code of for our CVPR 2018 paper "Neural Baby Talk"
https://arxiv.org/abs/1803.09845
MIT License
523 stars 123 forks source link

undefined symbol: _Py_Dealloc #22

Open imrankhurram opened 5 years ago

imrankhurram commented 5 years ago

I just want to run this on new images to see the results. I am using the evaluation section and getting this error:

ImportError: /home/nk/NeuralBabyTalk/pooling/roi_align/_ext/roi_align/_roi_align.so: undefined symbol: _Py_Dealloc

julycetc commented 5 years ago

i meet the same problem too!

julycetc commented 5 years ago

I just want to run this on new images to see the results. I am using the evaluation section and getting this error:

ImportError: /home/nk/NeuralBabyTalk/pooling/roi_align/_ext/roi_align/_roi_align.so: undefined symbol: _Py_Dealloc

you could run make.sh in the roi_align folder. Then, the problem is solved.

imrankhurram commented 5 years ago

I just want to run this on new images to see the results. I am using the evaluation section and getting this error: ImportError: /home/nk/NeuralBabyTalk/pooling/roi_align/_ext/roi_align/_roi_align.so: undefined symbol: _Py_Dealloc

you could run make.sh in the roi_align folder. Then, the problem is solved.

If I run make.sh then I get this error: cffi.error.VerificationError: LinkError: command 'x86_64-linux-gnu-gcc' failed with exit status 1

imrankhurram commented 5 years ago

another error written above in the logs is:

x86_64-linux-gnu-gcc: error: /home/ikhurram/neuralbabytalk/NeuralBabyTalk/poolin g/roi_align/src/roi_align_kernel.cu.o: No such file or directory

KaiserLew commented 5 years ago

I just want to run this on new images to see the results. I am using the evaluation section and getting this error: ImportError: /home/nk/NeuralBabyTalk/pooling/roi_align/_ext/roi_align/_roi_align.so: undefined symbol: _Py_Dealloc

you could run make.sh in the roi_align folder. Then, the problem is solved.

If I run make.sh then I get this error: cffi.error.VerificationError: LinkError: command 'x86_64-linux-gnu-gcc' failed with exit status 1

Hi, i met the same problem. Have you solved it?

imrankhurram commented 5 years ago

I just want to run this on new images to see the results. I am using the evaluation section and getting this error: ImportError: /home/nk/NeuralBabyTalk/pooling/roi_align/_ext/roi_align/_roi_align.so: undefined symbol: _Py_Dealloc

you could run make.sh in the roi_align folder. Then, the problem is solved.

If I run make.sh then I get this error: cffi.error.VerificationError: LinkError: command 'x86_64-linux-gnu-gcc' failed with exit status 1

Hi, i met the same problem. Have you solved it?

Unfortunately, No

gorgeousyouth commented 5 years ago

I just want to run this on new images to see the results. I am using the evaluation section and getting this error: ImportError: /home/nk/NeuralBabyTalk/pooling/roi_align/_ext/roi_align/_roi_align.so: undefined symbol: _Py_Dealloc

you could run make.sh in the roi_align folder. Then, the problem is solved.

If I run make.sh then I get this error: cffi.error.VerificationError: LinkError: command 'x86_64-linux-gnu-gcc' failed with exit status 1

Hi, i met the same problem. Have you solved it?

Unfortunately, No

I just want to run this on new images to see the results. I am using the evaluation section and getting this error: ImportError: /home/nk/NeuralBabyTalk/pooling/roi_align/_ext/roi_align/_roi_align.so: undefined symbol: _Py_Dealloc

you could run make.sh in the roi_align folder. Then, the problem is solved.

If I run make.sh then I get this error: cffi.error.VerificationError: LinkError: command 'x86_64-linux-gnu-gcc' failed with exit status 1

Hi, i met the same problem. Have you solved it?

Unfortunately, No

excuse me, have you solved the problem?

yongduek commented 5 years ago

Ubuntu 18.04.

The python version caused the problem in my case. In make.sh, python was changed to python2.7, then the command started running:

# python2.7 main.py --path_opt cfgs/normal_coco_res101.yml --batch_size 20 --cuda True --num_workers 20 --max_epo ch 30 --inference_only True --beam_size 3 --start_from save/coco_nbt_1024

The whole code must be built for python2.7, but sometimes the command python is linked to python3 instead of python2.7.

berthyf96 commented 5 years ago

A post in this issue helped me.

Before running pooling/roi_align/make.sh, add the following lines to pooling/roi_align/make.sh (right after line 10):

export PATH=/usr/local/cuda-8.0/bin${PATH:+:${PATH}}
export CPATH=/usr/local/cuda-8.0/include${CPATH:+:${CPATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

You might also want to uncomment line 7 (but this might not be necessary): export C_INCLUDE_PATH=/opt/cuda/include

After you've edited the shell script, you can do sh make.sh in pooling/roi_align. This fixed the ImportError for me.