Reputation: 1753
I have installed tensorflow 2.1.0 in a virtualenv on my home laptop (Ubuntu 18.04). When I import tensorflow as tf
from terminal I get the following warning:
2020-03-12 12:17:56.485098: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer.so.6'; dlerror: libnvinfer.so.6: cannot open shared object file: No such file or directory
2020-03-12 12:17:56.485179: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer_plugin.so.6'; dlerror: libnvinfer_plugin.so.6: cannot open shared object file: No such file or directory
2020-03-12 12:17:56.485189: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:30] Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
However, the import works and I can run code as usual.
But when I try to do the same in a Jupyter Notebook, I get error. import tensorflow as tf
returns:
ModuleNotFoundError: No module named 'tensorflow'
I have no GPU on my laptop, but tensorflow 2.1 should be able to run on either GPU and CPU-only machines. And in fact I installed it successfully at work with no problems. What can I do to fix this?
Upvotes: 1
Views: 696
Reputation:
You can suppress the Warnings message by using below code
import logging, os
logging.disable(logging.WARNING)
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
import tensorflow as tf
Regarding error on Jupyter notebook,
Open anaconda-navigator and create new environment and launch the jupyter-notebook. Now install tensorflow
!pip install tensorflow
Upvotes: 1