导入Tensorflow时Jupyter内核死亡，Jupyter Notebook出现问题

Question

I use Anaconda version 4.5.8 with Python 3.6.6. After installing tensorflow by using the following commands from the tensorflow website:

conda create -n tensorflow pip python=3.5
activate tensorflow
pip install --ignore-installed --upgrade tensorflow

I've done as they have instructed and yet my kernel dies whenever I import tensorflow as tf in my Jupyter notebook.

I've also looked at other solutions on this website and uninstalled all of tensorflow and installed pip install tensorflow==1.5 . No matter my kernel still dies.

I've created another new environment in Anaconda and installed pip install tensorflow==1.51 and still my kernel dies. I've downgraded to python2.7 , and my kernel dies.

I've downgraded my tensorflow and my python and the kernel still ends up dead. Can anyone find a solution to this?

Edit: Rechecked solutions again, others don't work. Edit: Tried environment changing as commenter suggested, didn't work.

Answer 1

I had the same problem (although it worked problem-free previously): I could do anything in Python as long as I didn't use tensorflow. Which is a bit difficult if you're trying to study convolutional networks. Installing tensorflow 1.5 (with pip3) instead of the current tensorflow did not help me either, but after trying conda instead of pip it worked like a charm!