inux version 5.15.0-58-generic (buildd#lcy02-amd64-101) (gcc (Ubuntu 11.3.0-1ubuntu1~22.04) 11.3.0, GNU ld (GNU Binutils for Ubuntu) 2.38)
conda version == 22.9.0
pip version == 22.2.2
python version 3.5 and 3.9
I have a machine with intel pentium g2020 cpu and it seems like it does not support AVX instructions. I am currently working on a project and need tensorflow (keras to be more precise) on my machine but any version that I install throws an error. I would need you help to find the suitable version for my machine and the a way to install it.
I tried different python version and tensorflow version but all of them throw errors like kernel died or your machine does not support AVX instructions.
I'm trying to install tensorflow,
at beginning I thought the error was caused by python 3.11
I changed the interpreter to python 3.10 but I still have the same problem.
I can't install tensorflow and the error is
ERROR: tensorflow_cpu-2.11.0-cp310-cp310-win_amd64.whl -
is not a supported wheel on this platform.'''
intel laptop running on win 11, amd ryzen 5000 series, amd radeon graphics
thanks
screen vscode
The API you are using basically a single version of tensorflow 2.11.0. Developers designed different wheels for differnt OS requiremnts. As as the latest version of tensorflow is 2.11.0 so don't wory installing through Above API. Install directly using PIP. Let PIP decide supported version.
pip install tensorflow
I want to install PyTorch on a Debian Linux-based operating system (ev3dev).
System specification (Lego Mindstorms EV3):
64 MB RAM
TI Sitara AM18080 (ARM926EJ-S core) #300 MHz
Installed Python versions: Python 3.5.3, Python 3.9.4
32 GB SD-Card
The following installation command:
pip3 install torch==1.10.0+cpu
leads to the error: Could not find a version that satisfies the requirement torch==1.10.0+cpu
Now, my question is, if it is possible to install and run PyTorch on such a system. And if it is possible to install, is it possible to run with acceptable performance?
Thank, you for any hints.
I am ecstatic that I have found a setup where I can train and test my custom object detection models but I am confused why the latest tensorflow-gpu version doesn't work on my PC when another setup works on my laptop:
The Setup I am using(PC):
OS: Windows 10
Python: 3.6.4
TensorFlow: 1.14.0(gpu)
CUDA: 10
cudnn: 9.0?(can't remember)
repo + location: https://github.com/tensorflow/models/research/object_detection
The Setup I am using(laptop):
OS: OSX
Python: 3.7.5
TensorFlow: 2.1.0-rc0(standard)
repo + location: https://github.com/tensorflow/models/research/object_detection
how(when?) can I train custom models using tensorflow-gpu v2.0.0 on my PC?
thx
If you need to run tensorflow-gpu on Windows, Microsoft Visual Studio is a pre-requisite. This requirement is also mentioned here. You can download the community version of Visual Studio 2019 here.
Though this requirement is not directly mentioned in tensorflow-gpu installation docs, it is a part of CUDA installation guide for Windows. Also you need to ensure you have a NVIDIA GPU card with CUDA Compute Capability 3.5 or higher. You can check if your GPU meets the Compute Capability requirement here.
Here are a few resources to help with tf-gpu installation:
https://towardsdatascience.com/installing-tensorflow-with-cuda-cudnn-and-gpu-support-on-windows-10-60693e46e781
https://www.youtube.com/watch?v=KZFn0dvPZUQ
I'm using Python 3.7, and it seems every solution I searched is on Conda environment, which is, I don't use.
My computer spec is intel i7 with nvidia gtx1050, and when I used the command "pip3 install --upgrade tensorflow-gpu" in command line, it always fires this error.
C:\WINDOWS\system32>pip3 install --upgrade tensorflow-gpu
Collecting tensorflow-gpu
Could not find a version that satisfies the requirement tensorflow-gpu (from versions: )
No matching distribution found for tensorflow-gpu
Anyone can help? I have CUDA Toolkit 9.2 installed, and the latest cuDNN installed. I'm running Windows 10.
When I search “TensorFlow” in https://pypi.org/project/tensorflow/#files
I find the following information:
tensorflow-1.8.0-cp36-cp36m-win_amd64.whl Wheel cp36 Apr 28, 2018
Till 2018-07-09, only python(<=3.6) could find a version that satisfies the requirement tensorflow.
I think if you want to use TensorFlow, maybe you should lower you Python version.
I found a version that satisfies the requirement tensorflow.
And I write down the method in the another answer on this page.
2018-07-30
I found tensorflow-1.9 for Python3.7 (tensorflow‑1.9.0‑cp37‑cp37m‑win_amd64.whl)
https://www.lfd.uci.edu/~gohlke/pythonlibs/
https://www.lfd.uci.edu/~gohlke/pythonlibs/#tensorflow
You would find this on the website:
TensorFlow, computation using data flow graphs for scalable machine learning.
Requires numpy+mkl and protobuf. The CUDA builds require CUDA 9.2 and CUDNN 9.2.
tensorflow‑1.9.0‑cp36‑cp36m‑win_amd64.whl
tensorflow‑1.9.0‑cp37‑cp37m‑win_amd64.whl
other version of TensorFlow.whl
https://github.com/fo40225/tensorflow-windows-wheel
Tensorflow is support only on Python 3.5.X versions. Try installing on any of these versions. This should fix the problem.
Just as a heads up, Tensorflow GPU WONT WORK with CUDA toolkit v9.2, you need to download and install v9.0. It's available by clicking 'legacy downloads' on the Cuda download page. TF looks for a specific file (cudart64_90.dll) that is from 9.0 and won't accept the equivalent version of the library (cudart64_92.dll) from the latest 9.2 version. The good news is that you DO NOT have to uninstall v9.2 of the toolkit, the Nvidia installer will just add a separate v9.0 folder in the Nvidia Cuda directory right next to the v9.2 folder you already have. I hope this helps.