libtensorflow_framework.so.2: cannot open shared object file: No such file or directory

[LeiLiu8102]

Hi, I have an error when installing the python interface of deepmd-kit.

I followed the instructions to install python interface of deepmd-kit and tensorflow:
python3 -m venv --system-site-packages tf1.14
source tf1.14/bin/activate
pip install --upgrade pip
pip install --upgrade tensorflow
pip install --upgrade tensorflow==1.14
pip install deepmd-kit

I verified my tensorflow installation by python -c "import tensorflow as tf;print(tf.reduce_sum(tf.random.normal([1000, 1000])))" and it outputs

Tensor("Sum:0", shape=(), dtype=float32)

When I execute the command dp -h,
it says

tensorflow.python.framework.errors_impl.NotFoundError: libtensorflow_framework.so.2: cannot open shared object file: No such file or directory

Could you give any clue what the problem might be? Thank you.

I tried to address this issue by linking libtensorflow_framework.so.2 to libtensorflow_framework.1, because I can find libtensorflow_framework.so.1 in tf1.14/lib64/python3.6/site-packages/tensorflow. However it doesn't work.

System info:
python 3.6.8
gcc 4.8.5
cmake 3.11.0
tensorflow 1.14.0
deepmd 1.2.0

[njzjz]

Please upgrade Tensorflow to 2.1.

[LeiLiu8102]

Thank you for your comment. I upgrade Tensorflow to 2.1, then executed dp -h, and found the error

Traceback (most recent call last):
File "/home/whuan104/lliu147/tfvenv/bin/dp", line 5, in
from deepmd.main import main
File "/home/whuan104/lliu147/tfvenv/lib64/python3.6/site-packages/deepmd/init.py", line 1, in
from .env import set_mkl
File "/home/whuan104/lliu147/tfvenv/lib64/python3.6/site-packages/deepmd/env.py", line 61, in
op_module = get_module("libop_abi")
File "/home/whuan104/lliu147/tfvenv/lib64/python3.6/site-packages/deepmd/env.py", line 58, in get_module
module = tf.load_op_library(module_path + "{}.{}".format(module_name, ext))
File "/home/whuan104/lliu147/tfvenv/lib64/python3.6/site-packages/tensorflow_core/python/framework/load_library.py", line 57, in load_op_library
lib_handle = py_tf.TF_LoadLibrary(library_filename)
tensorflow.python.framework.errors_impl.NotFoundError: /home/whuan104/lliu147/tfvenv/lib/python3.6/site-packages/deepmd/libop_abi.so: undefined symbol: _ZN10tensorflow8OpKernel11TraceStringEPNS_15OpKernelContextEb

Could you please let me the possible reason? Thank you.

[njzjz]

I don't know why. It was working in Google Colaboratory. You can try pip install git+https://github.com/deepmodeling/deepmd-kit to reinstall deepmd-kit.