... is not a supported wheel on this platform

Getting Started
,
1

Hey.

I have a wheel issue as I have seen a couple of others have but after going through all the different threads I can't seem to get a solution going. Quick info:

I run windows 10 and have created a virtual env via conda. My env specs are:

Python 3.6.9
Pip v. 19.3.1

When I go print(wheel.pep425tags.get_supported())

I get:

[('cp36', 'cp36m', 'win_amd64'), ('cp36', 'none', 'win_amd64'), ('cp36', 'none', 'any'), ('cp3', 'none', 'any'), ('cp35', 'none', 'any'), ('cp34', 'none', 'any'), ('cp33', 'none', 'any'), ('cp32', 'none', 'any'), ('cp31', 'none', 'any'), ('cp30', 'none', 'any'), ('py3', 'none', 'win_amd64'), ('py36', 'none', 'any'), ('py3', 'none', 'any'), ('py35', 'none', 'any'), ('py34', 'none', 'any'), ('py33', 'none', 'any'), ('py32', 'none', 'any'), ('py31', 'none', 'any'), ('py30', 'none', 'any')]

My file is located at (and is named) "C:\Users\Admin\Desktop\prodigy-1.9.5-cp36.cp37.cp38-cp36m.cp37m.cp38-win_amd64.whl"

I have run the following command in bot pip and pip3 (and also tried adding and removing m at the last cp38):

pip install prodigy-1.9.5-cp36.cp37.cp38-cp36m.cp37m.cp38-win_amd64.whl

It still says: "is not a supported wheel on this platform."

Any advice?

2

Hi! What happens if you rename the file from ...cp38-win_amd64.whl to ...cp38m-win_amd64.whl? It's a wild guess, but there has been an issue with the setuptools appending/not appending the m flag (see here), which means we had to rename all wheels we ship for our packages. But depending on the version of pip, maybe it does care about the m being there in your case.

3

It says the same thing. Oddly, by mistake, I pip'ed it into my base env and that worked but it still doesn't work on the env I dedicated to it. Are there any dependencies I need to make sure exists first? I haven't gone through the list but I have all sorts of weird stuff in my base env.

4

That's very strange – but also kinda good news :slightly_smiling_face: The dependencies shouldn't make a difference (and the wheel should install everything it needs). The error you're seeing means that pip doesn't even acknowledge the file as valid, so you're not even getting to the install stage.

Which pip versions are you running in both envs? And if they're different, maybe try installing the same pip version your base env has in your new env? It could easily be some pip version weirdness.

5

You were right on the money - different versions of pip. Thank you for your help and a really amazing tool. I can't wait to play around with it!

1 Like
6

@rolflund Glad you got it working! Just curious, which versions of pip were you using, and which one worked? Just so I know in case this (or a similar problem) comes up again in the future :slightly_smiling_face:

7

Hi, I also get the same error. I'm using ubuntu OS with pip version 19.0.3. May I know which version required, please.

8

The version shouldn't matter – just make sure the wheel is named correctly and matches your platform, because that's what pip cares about. Also make sure that the local path is correct (and try pip install *.whl so you don't have to repeat the file name on the command line).

9

hi i'm also getting same issue plz help me correcting it.
i want to install pyaudio but it is coming *whl is not supported.......

1 Like
10

Hello. I am having the same problem on Windows 10. I am not using an environment to install. I've tried the above, without much avail. Would anyone have any suggestions?

11

Hi! Which version of Python are you using? Did you double-check that the file definitely exists, and that you're using the Windows installer? Does pip install *.whl help?

12

Ines. Thank you. It is all sorted. I think it is the same issue that other have encountered - it got all sorted out when I installed inside a conda environment. Congratulations - really neat software.

1 Like
13

Hi Ines. I seem to be having the same issue.

I'm trying to install within a conda env.

Here are some background details:

  • The filename downloaded is: prodigy-1.10.8-cp36.cp37.cp38-cp36m.cp37m.cp38-win_amd64.whl
  • The OS is win-amd64
  • pip version is 21.0.1
  • Python version is 3.9.4

Thanks in advance for any help.
Darren

14

Hi! This is likely the problem – we currently only have pre-built wheels for 3.6, 3.7 and 3.8 (wheels for 3.9 are coming in the next release). So the easiest solution would be to just use a separate virtual environment with 3.8 and it should work :slightly_smiling_face: (Highly recommend pyenv btw if you're not already using it – it also has a Windows version!)

2 Likes
15

Amazing - thanks Ines :pray:

1 Like
16

Is this still the case? I'm currently getting the dreaded "prodigy-1.10.8-cp36.cp37.cp38-cp36m.cp37m.cp38-win_amd64.whl is not a supported wheel on this platform" error running on 3.9.6 64-bit on a Windows 10 machine. PIP version is up to date (21.1.3). Being new to Python, going through the process of installing an old Python can be a "process" (no binary installers), especially if it doesn't fix the problem.

Thanks!

17

Prodigy v1.10.x only ships with wheels for 3.6 to 3.8, but the upcoming v1.11 (currently available as a nightly pre-release) will have wheels for 3.9, as well as additional wheels for ARM architectures.

If you do want to install an older version of Python, I'd recommend using pyenv, which lets you manage multiple versions without messing with your system Python.

1 Like
18

A heads up to any new Python people here who have my setup—I'm using vscode and Anaconda. I previously installed Python 3.9 separately on my machine and was using it in VScode for my projects. In order to get Prodigy to install, you'll need v3.8 (as ines indicated), so Conda is a great solution to this as it comes with v3.8.

VSCODE wasn't listening to my Conda interpreter (which is 3.8) selection to install prodigy. Even though my 3.8 interpreter was selected, when I typed "python --version" in the terminal, Python 3.9 showed up. If you currently have Python 3.9 installed AND Conda, you may need to uninstall Python 3.9 (if you don't absolutely need that version). This should circumvent the "needing to downgrade" problem without having to use pyenv.

1 Like