I've installed Ubuntu 20.04 on a Raspberry Pi. Both my Ubuntu and Python3 installations are 64-bit. I'm now trying to install Tensorflow so that I can run (not train) a model.

I've tried various blog-posts, installed all the dependencies, etc. However, I am always faced with the same error:

> pip3 install tensorflow
ERROR: Could not find a version that satisfies the requirement tensorflow (from versions: none)
ERROR: No matching distribution found for tensorflow

and running pip install --upgrade tensorflow gives the same error.

What is wrong? How can I install tensorflow on my Raspberry Pi?

Background of my questions

I have an old Lenovo ThinkPad W530 laptop with NVIDIA Corporation GK107GLM [Quadro K1000M] / Quadro K1000M/PCIe/SSE2 GPU which is with compute compatibility 3.0 and is supporting CUDA. I want to build some TensorFlow examples and project on Ubuntu for my Udacity nanodegree.

I just installed Ubuntu 20.04 and upgraded my NVIDIA drivers so when I execute nvidia-smi it gives me the following result:

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 450.51.05    Driver Version: 450.51.05    CUDA Version: 11.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Quadro K1000M       On   | 00000000:01:00.0 Off |                  N/A |
| N/A   59C    P0    N/A /  N/A |    496MiB /  1999MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                              
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|    0   N/A  N/A      5470      G   /usr/lib/xorg/Xorg                 39MiB |
|    0   N/A  N/A      7504      G   /usr/lib/xorg/Xorg                146MiB |
|    0   N/A  N/A      7772      G   /usr/bin/gnome-shell              147MiB |
|    0   N/A  N/A     34770      G   /proc/self/exe                    158MiB |
+-----------------------------------------------------------------------------+

As I am trying to build the TensorFlow Code using GPU support, I found out that I need to install CUDA and follow this article about how to install latest version of CUDA 11.0 on Ubuntu 20.04. I went through the whole path and then I found out two problems:

1. My GPU is only support Compute Compatibility 3.0 while the minimum support for TensorFlow is 3.5
2. To overcome problem 1, I have to build the tensor flow from sources to get a build that supports CUDA 3.0 and to do that I have to use CUDA 10.1 which means I have to start from the beginning.

I removed CUDA 11.0 following the CUDA documentation. I started to install CUDA 10.1 on Ubuntu 20.04 and now I got confused.

CUDA 10.1 installation documentation process is about installing CUDA 10.1 from a .deb file for Ubuntu 18.04 and there is no support for CUDA 10.1 on Ubuntu 20.04.

So I started to follow the same steps of the article but for CUDA 10.1 instead CUDA 11. I was able to install cuda 10.1 using the command:

sudo apt-get -f install cuda-toolkit-10-1 cuda-libraries-10-1

so that I don't override my installed NVIDIA driver 450 (I did that after several trials).

When I tried to execute command

sudo apt install nvidia-cuda-toolkit

It gave me errors due to dependency problems. To overcome these errors I searched for a solution and then I came across this article which is recommending that:

For the sake of being verbose, do not try to use 18.10 or 18.04 CUDA 10.1 for Ubuntu 20.04. I learned that the hard way, lol!

So, you can install CUDA 10.1 in Ubuntu 20.04 by running,

$ sudo apt install nvidia-cuda-toolkit

I uninstalled CUDA Toolkit 10.1 that I installed using CUDA documentation and I followed this recommendation and the command passed successfully. But now I am not sure if this is it, or am I missing something?!

When I tried to know the different between both ways I found out that the source of nvidia-cuda-toolkit using the command

apt-cache policy nvidia-cuda-toolkit

it gives me the following result that indicates that the source is the Ubuntu repositories

nvidia-cuda-toolkit:
  Installed: 10.1.243-3
  Candidate: 10.1.243-3
  Version table:
 *** 10.1.243-3 500
        500 http://eg.archive.ubuntu.com/ubuntu focal/multiverse amd64 Packages
        100 /var/lib/dpkg/status

My Questions

1. If the official installation documentation of CUDA 10.1 for Ubuntu 18.04 as well as CUDA 11 for Ubuntu 20.04 doesn't indicate any thing regarding installing the package of nvidia-cuda-toolkit so what is the difference between nvidia-cuda-toolkit and cuda-toolkit-10-1?
2. Is it enough as indicated by the article to only install nvidia-cuda-toolkit if I am trying to use CUDA toolkit 10.1?
3. Is CUDA 10.1 runtime libraries are now installed? If not, How can I install them right now? should I remove the nvidia-cuda-toolkit or can I follow the steps from the documentation?
4. How can I validate my installation?
5. When I tried to build CUDA Samples, didn't find the cuda-install-samples-10.1.sh file in my installation. How can I install the samples?

I have been working on trying to build my examples and so far I failed to do that, so I really appreciate your support for this.

I am a linux newcomer trying to set up tensorflow with GPU for Python on Pop!_os 20.04. I tried using the proposed Pop!_os apt packages for this but they didn't work for me.

After a lot of struggle on my own and possible apt pollution, I managed to install Tensorflow with CUDA and cuDNN by following this answer https://askubuntu.com/a/1251052

Although Python's tensorflow seems to properly recognize my GPU, right after importing the library I get this weird message:

Python 3.8.2 (default, Apr 27 2020, 15:53:34) 
[GCC 9.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
Invalid MIT-MAGIC-COOKIE-1 key
>>>

Can someone tell me what this means or give me some clue on what should I do (maybe I can take that as a warning and live with it), please?

Not sure if this qualifies as an Ubuntu question but since I followed an Ubuntu guide to reach this point and Pop is Ubuntu-based it just felt right to ask here.

System info (advice on what to include/not include here next time is appreciated):

OS: Pop!_OS 20.04 LTS x86_64
Host: ERAZER X6807 MD61178 Standard 
Kernel: 5.4.0-7634-generic
Packages: 2407 (dpkg)
Shell: bash 5.0.16
Resolution: 1920x1080 
DE: GNOME
WM: Mutter
Terminal: gnome-terminal 
CPU: Intel i7-8750H (12) @ 4.100GHz
GPU: Intel UHD Graphics 630
GPU: NVIDIA GeForce GTX 1060 Mobile
Memory: 15868MiB

Thanks in advance!

Is there any way to install tensorflow-gpu on WSL2?
After trying to set up CUDA on WSL and running nvidia-smi, it had the error:

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

If anyone knows how to fix it, that would be appreciated.

I just upgraded my system to Ubuntu 19.04 and realized Cuda 10.0 is not getting installed. tensorflow is still not built for Cuda 10.1 and it would be great if I can downgrade my Nvidia driver from 418 to 410 if possible. I am using Geforce GTX 1060.