This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(208, 68%, 30%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EMJ%3C/text%3E%3C/svg%3E)
I am using azure ML to run an python code in a jupyter notebook using tensorflow. I'm using "Python 3.6 - Azure ML" and an instance of azure compute to run it. It's showing tensorflow version as 2.1.0 .When I run the same code using VS Code connected to azure portal remotely, it's shows tensorflow version as 2.5.0. Why this discrepency?
Also I need tensorflow version 2.3.0 or up but there is not way for me to upgrade that using azure ml portal. Any help will be appreciated. Thank you!
@M J I think the discrepancy is because the notebook is using a different environment with the kernel Python 3.6 - Azure ML. You can install the required version of tensorflow or any package with pip install from a new cell in the notebook.
!pip install tensorflow==2.3.0
This should help you use this environment from the notebook. Thanks.
I don't think they are using different enviroment. It is same "Python 3.6 - Azure ML" located at the same place in the VM. There is something wrong with the way one of these work.
See the highlighted in above screenshots.
The suggestion didn't work. When I upgrade to tensorflow 2.3 and try to print the version, it still stays the same 2.1
@M J I think you are importing the previous version as tf and then installing the required version and printing the version of the previously imported variable.
Please try to import tensorflow as tf1 in a new cell and then print the version or you could simply restart your kernel and check the version by commenting out the install command in your cell.
It still didn't work. The same thing works when I do it from VS Code. What's the difference between these two?
