pointnet2-tensorflow2 icon indicating copy to clipboard operation
pointnet2-tensorflow2 copied to clipboard

Error running the code with cuda 11.8

Open FSet89 opened this issue 4 months ago • 0 comments

I tried to compile tf_ops with cuda 11.8 and -std=c++17. When I run the train code I get this error. How can I fix it?

Traceback (most recent call last): File "pointnet2-tensorflow2/train_scannet.py", line 12, in from models.sem_seg_model import SEM_SEG_Model File "pointnet2-tensorflow2/./models/sem_seg_model.py", line 10, in from pnet2_layers.layers import Pointnet_SA, Pointnet_FP File "pointnet2-tensorflow2/./pnet2_layers/layers.py", line 4, in from . import utils File "pointnet2-tensorflow2/./pnet2_layers/utils.py", line 8, in from .cpp_modules import ( File "pointnet2-tensorflow2/./pnet2_layers/cpp_modules.py", line 9, in sampling_module=tf.load_op_library('./tf_ops/sampling/tf_sampling_so.so') File "/home/francesco/.local/lib/python3.10/site-packages/tensorflow/python/framework/load_library.py", line 54, in load_op_library lib_handle = py_tf.TF_LoadLibrary(library_filename) tensorflow.python.framework.errors_impl.NotFoundError: ./tf_ops/sampling/tf_sampling_so.so: undefined symbol: _ZN10tensorflow14kernel_factory17OpKernelRegistrar12InitInternalEPKNS_9KernelDefESt17basic_string_viewIcSt11char_traitsIcEESt10unique_ptrINS0_15OpKernelFactoryESt14default_deleteISA_EE

FSet89 avatar Feb 19 '24 12:02 FSet89