Is the onnxconverter-common library being deprecated? If so, is there another library which provides similar functionality which is going to be maintained long-term? #7022
Replies: 1 comment 2 replies
-
|
I don't think it is being actively maintained. Could you share your use case? If float16 conversion is the main use case, we can create something in |
Beta Was this translation helpful? Give feedback.
-
|
Thanks, that is useful to know. I am currently just using the |
Beta Was this translation helpful? Give feedback.
-
|
@justinchuby I have opened a PR which adds the code for float16 conversion to the ir-py library as you suggested: onnx/ir-py#86 |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
I have been using the onnxconverter-common library (https://github.com/microsoft/onnxconverter-common) to convert deep learning models to a 16-bit float version following the instructions described here in the onnx runtime documentation.
It worked perfectly when I was testing locally with python3.12, but it wasn't until my CI/CD ran that I discovered that it fails with python3.13. I then took a closer look at the onnx-common repo and noticed that there hasn't been an official release of the package since August 2023, and only very sporadic updates to the source code since.
Are people using something different instead of onnxconverter-common now? Or have the maintainers of the package just moved on to other things?
Beta Was this translation helpful? Give feedback.
All reactions