How to debug pip installation error: No matching distribution found for onnxruntime

[snnn]

If when you tried to install onnxruntime python package pip said: "No matching distribution found for onnxruntime", here is how to debug it. Please follow the instructions to get more information before open an issue in Github. The title of your Github issue needs be more accurate instead of just saying "No matching distribution found for onnxruntime". If you need our help to figure out why there was "no matching", please finish at least the step 1 of below and give us:

1. The "sys.version" information
2. The list of "Compatible tags"

Step 1: find out the pip tags that your python supports

Please open a command prompt and run

pip debug --verbose 

It will output a lot of information.

First , at the very beginning of the output you could find your python version. Search the line starts with "sys.version". For example, you may see things like:

sys.version: 3.13.2 (tags/v3.13.2:4f8bb39, Feb  4 2025, 15:23:48) [MSC v.1942 64 bit (AMD64)]

It means your python version is 3.13.2. The major version is 3. The minor version is 13. They are the key information we are looking for.

Then, please examine the output's last section: "Compatible tags:" . As our python version is 3.13, we will only look for the lines start with cp313.

In this example, I have:

Compatible tags: 45
  cp313-cp313-win_amd64
  cp313-abi3-win_amd64
  cp313-none-win_amd64

If your operating system is Linux, you may filter the output further by using grep. You may first run:

$ pip debug --verbose |grep sys.version

to get the python version. Now let's assume it is 3.9. Then please run

$ pip debug --verbose |grep cp39 |grep abi3-manylinux

It will list the tags your python supports. You may see things like:

  cp39-abi3-manylinux_2_35_aarch64
  cp39-abi3-manylinux_2_34_aarch64
  cp39-abi3-manylinux_2_33_aarch64
  cp39-abi3-manylinux_2_32_aarch64
  cp39-abi3-manylinux_2_31_aarch64
  cp39-abi3-manylinux_2_30_aarch64
  cp39-abi3-manylinux_2_29_aarch64
  cp39-abi3-manylinux_2_28_aarch64
  cp39-abi3-manylinux_2_27_aarch64
  cp39-abi3-manylinux_2_26_aarch64
  cp39-abi3-manylinux_2_25_aarch64
  cp39-abi3-manylinux_2_24_aarch64
  cp39-abi3-manylinux_2_23_aarch64
  cp39-abi3-manylinux_2_22_aarch64
  cp39-abi3-manylinux_2_21_aarch64
  cp39-abi3-manylinux_2_20_aarch64
  cp39-abi3-manylinux_2_19_aarch64
  cp39-abi3-manylinux_2_18_aarch64
  cp39-abi3-manylinux_2_17_aarch64
  cp39-abi3-manylinux2014_aarch64

Please take a note of this information for later use.

Step 2: use pypi.org to find out which pip tags ONNX Runtime provides

Open pypi.org in a web browser, find the package you wanted to install, then click "Download files". You will see a page like this:

Then look at the "Built Distributions" section. Each file has a tag.
For example, if the file name is : "onnxruntime-1.10.0-cp39-cp39-win_amd64.whl"
The string can be split into three parts:
The first part, "onnxruntime", is the package name.
The second part, "1.10.0" , is the version.
The last part, "cp39-cp39-win_amd64" is the pip tag. Does your python support this tag? (see step 1)

So, when you see "No matching distribution found for onnxruntime", it means that there is a set of pip tags that you python can support and there is a set of pip tags that onnx runtime team published to pypi but the two set have nothing in common! (The intersection of two sets is empty).

Normally, the reason could be:

1. Your python version is too new or too new. For example, onnxruntime 1.10 has tags of cp39*, cp38*, cp37*, cp36*, but if your python version is not in the [3.6, 3,9] range, it won't match.
2. Your glibc version is too old. For example, if your pip supports manylinux_2_17, ..., manylinux_2_34, but the ONNX Runtime python package's tag is manylinux_2_35, then it means the package requires glibc 2.35+. Then you need to upgrade your operating system to a new major version.
3. Your CPU arch is not supported.

If you cannot figure out the reason by yourself, you may post the output of your "pip debug --verbose" command to our Github issues, then we will check it for you.

Additional steps

You may pip install command with a fixed onnxruntime version and with flags "-vv". Like this:

python3 -m pip install -vv onnxruntime==1.16.3

If you use a custom feed, it would be something like this:

python3 -m pip install -vv --index-url=https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ort-cuda-12-nightly/pypi/simple/ onnxruntime==1.16.3

The command above will output a lot of things. You can redirect the log to a file and search "Skipping link: none of the wheel's tags". It explains why the installations failed. You may post the log to our Github issues if you need our help to examine to result.

Other Notes

If your pip command has "--index-url" that points our Azure DevOPS feed, you may need to install all the dependencies of the onnxruntime package first. You may need to get the dependencies from the official pypi.org. For example, to install a nightly package from our custom Azure DevOps feed, you may need to two commands:

python -m pip install coloredlogs flatbuffers numpy packaging protobuf sympy
python -m pip install -i https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ORT-Nightly/pypi/simple/ ort_nightly

Thanks.