search

MVIG-SJTU / pointSIFT

a module for 3D semantic segmentation in point clouds.
644 stars 119 forks source link

tensorflow.python.framework.errors_impl.NotFoundError: /home/cc/pointSIFT-master/tf_utils/tf_ops/pointSIFT_op/tf_pointSIFT_so.so: undefined symbol: _ZN10tensorflow8internal21CheckOpMessageBuilder9NewStringEv #10

Closed ZeweiXu closed 6 years ago

ZeweiXu commented 6 years ago

I am using python2.7, Tensorflow1.4 and cuda9.1. I successfully complied the files in the four folders: sampling, pointSIFT_op, interpolation, and grouping. But when I run the training: python train_and_eval_scannet.py, the below error always pops up. Any help would be greatly appreciated!!!

/home/cc/anaconda2/lib/python2.7/site-packages/h5py/init.py:36: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type. from ._conv import register_converters as _register_converters Traceback (most recent call last): File "train_and_eval_scannet.py", line 14, in import pointSIFT_pointnet as SEG_MODEL File "/home/cc/pointSIFT-master/models/pointSIFT_pointnet.py", line 6, in from pointSIFT_util import pointSIFT_module, pointSIFT_res_module, pointnet_fp_module, pointnet_sa_module File "/home/cc/pointSIFT-master/tf_utils/pointSIFT_util.py", line 12, in from pointSIFT_op import pointSIFT_select, pointSIFT_select_four File "/home/cc/pointSIFT-master/tf_utils/tf_ops/pointSIFT_op/pointSIFT_op.py", line 14, in pointSIFT_module = tf.load_op_library(os.path.join(BASE_DIR, 'tf_pointSIFT_so.so')) File "/home/cc/anaconda2/lib/python2.7/site-packages/tensorflow/python/framework/load_library.py", line 56, in load_op_library lib_handle = py_tf.TF_LoadLibrary(library_filename, status) File "/home/cc/anaconda2/lib/python2.7/site-packages/tensorflow/python/framework/errors_impl.py", line 473, in exit c_api.TF_GetCode(self.status.status)) tensorflow.python.framework.errors_impl.NotFoundError: /home/cc/pointSIFT-master/tf_utils/tf_ops/pointSIFT_op/tf_pointSIFT_so.so: undefined symbol: _ZN10tensorflow8internal21CheckOpMessageBuilder9NewStringEv

jmydurant commented 6 years ago

Make sure:

  1. path in compile file must fit the python you used (most likely). Double check your python env or anaconda path.
  2. strongly recommend python3. For Tensorflow 1.4, you may not use CuDNN 9.1. Because it's too new for this tf version.
chencidhao226342 commented 6 years ago

python3.6 and tensorflow1.8 cuda8.0, but it still has this problem, i have checked the path. what should I do? thanks!

ZeweiXu commented 6 years ago

Thank you for the answer, I reinstalled and use the exactly same Cuda, cudnn and tensorflow versions mentioned by the authors and it works.

ashish-kubade commented 5 years ago

Hi Team, I am getting above error, and the solutions mentioned above, could not resolve my problem, hence I am posting here again. I have checked the python, cuda and cudnn versions, which are python=3.5, cuda = 8.0, CuDNN = 7.0. Could anyone please suggest me any workarounds if above configuration is valid.

TIA :)

bodhicheng commented 5 years ago

The problem is caused by the difference between the local and compiled gcc versions of tensorflow. Conda's tensorflow is compiled by gcc7. My local version is 4.8. When I installed with pip, the problem was solved.

dennisushi commented 1 year ago

A similar error happened to me - I had to fix by reinstalling tensorflow-addons only. Reinstalling anything else did not help.