"wheel not supported" for pip install

Could you try to update conda and and install PyTorch again?

conda update -n base conda

@chen2018 we do not provide binary packages for IBM PPC64 machines (which you are using). To install PyTorch on ppc64, see https://developer.ibm.com/code/howtos/install-pytorch-on-power

Hi,

I’m getting this error:
torch-0.4.1-cp37-cp37m-win_amd64.whl is not a supported wheel on this platform.

1 Like

Hi,

I’ve read the comments above, currently the only version of pytorch that I can install is the with with “-none-” in the filename, which is 0.1.12. Other versions give the “wheel not supported” error.

Can you either solve this error, or update the last pytorch versions also to include the “-none-” in the filename?

Thanks!

I’ve just succeeded to install newer versions by updating python to 3.6.
My previous python version were 2.7 and 3.4 (ubuntu 14).
However, can you please update the pytorch versions also for python2.7?

I have a similar issue on windows 10 .
error: torch-1.1.0-cp37-cp37m-win_amd64.whl is not a supported wheel on this platform.

pip3 install https://download.pytorch.org/whl/cu100/torch-1.1.0-cp37-cp37m-win_amd64.whl

i have python 3.7 and pip is updated too and have CUDA10

You may have 32-bit Python or Windows installed.

I seem to have the same error on Ubuntu 19.04 64 bits with Cuda 10 and CudNN 7.6 ?

this is my config for installing pytorch on odroid board

uname -a  :
Linux odroid64 3.14.79-116 #1 SMP PREEMPT Tue Sep 26 01:19:06 BRT 2017 aarch64 aarch64 aarch64 GNU/Linux

and

lsb_release -a
No LSB modules are available.
Distributor ID:	Ubuntu
Description:	Ubuntu 16.04.3 LTS
Release:	16.04
Codename:	xenial

why i couldn’t install pytorch using pip?
python 3.6 no cuda
@smth


Here I am also facing the same issue can anyone help

PyTorch wheels do not support the RaspberryPi platform and you would need to build from source or search for user-provided binaries.

@ptrblck could you please explain me in detail?

You won’t be able to use something like pip install torch as the pre-built binaries do not support your platform.
Instead you would need to build PyTorch from source as explained here.