Could not find a version that satisfies the requirement tensorflow-federated==0.52.0

[SamuelGong]

As of today, the released version of tensorflow-federated is 0.52.0, which can be found in pypi and other occasions like Google Colab. However, when I executed

pip install tensorflow-federated==0.52.0

in my local machine, errors go like

ERROR: Could not find a version that satisfies the requirement tensorflow-federated==0.52.0 (from versions: 0.1.0, 0.2.0, 0.3.0, 0.4.0, 0.5.0, 0.6.0, 0.7.0, 0.9.0, 0.10.0, 0.10.1, 0.11.0, 0.12.0, 0.13.0, 0.13.1, 0.14.0, 0.15.0, 0.16.0, 0.16.1, 0.17.0, 0.18.0, 0.19.0, 0.20.0, 0.21.0, 0.22.0, 0.23.0, 0.24.0, 0.26.0, 0.27.0, 0.28.0, 0.29.0, 0.30.0, 0.31.0, 0.32.0, 0.33.0, 0.34.0, 0.36.0, 0.37.0, 0.38.0, 0.39.0, 0.40.0, 0.41.0, 0.42.0, 0.43.0, 0.44.0, 0.45.0, 0.46.0, 0.47.0, 0.48.0)
ERROR: No matching distribution found for tensorflow-federated==0.52.0

For your information, my python version is 3.9.16 (main, Mar 8 2023, 04:29:44) [Clang 14.0.6 ] with the up-to-date pip 23.0.1, and I was just using the default global.index-url for pip, i.e. when I executed the command

pip config list

it outputs nothing.

Thus, I wonder what tricks I should look for before I can enjoy the latest tensorflow-federated. This would help me run the Colab tutorial notebook on my local machine, avoiding taking up the public CPU and memory.

Really looking forward to your help!

[SamuelGong]

Still not resolved. Seems to me that the issue occurred in CentOS 7 and MacOS, while everything went well in Ubuntu. Is it possible that OS matters?

[zcharles8]

Ah, are you using Mac? Unfortunately we stopped supporting Mac, so this is likely why you are not able to find the newest versions of TFF.

[SamuelGong]

Sure I was using Mac. I see.

On the other hand, I also used CentOS 7 but failed. Does TFF support CentOS 7 any more?

[SamuelGong]

I even found that under Ubuntu 18.04 TFF is at most 0.48.0, while for Ubuntu 22.04 I can have 0.52.0. While I could not understand the reason behind, does that mean TFF only supports a limited release of OSes?

[huili0140]

This sounds like a question about the python package rather than a "federated learning" question to us. When we had the python executor we supported any version of linux, after we started releasing the C++ executor the linux compatibility was reduced, which we fixed by reducing the linux compatibility.

You can see that at https://github.com/tensorflow/federated/blob/main/tensorflow_federated/tools/python_package/build_python_package.sh#L92

We believe this isn't necessarily a CentOS 7 issue either.

[SamuelGong]

I checked the package provided from pypi.org, the whl file of 0.53.0 is

  tensorflow_federated-0.53.0-py3-none-manylinux_2_31_x86_64.whl

while the whl file of 0.48.0 (which can be installed on my CentOS 7) is

  tensorflow_federated-0.48.0-py2.py3-none-any.whl

Since version 0.49.0, the file use "manylinux_2_31_x86_64" instead of "any" any more. Do you think this change restricted pip to install version higher than 0.48.0 on CentOS 7?