-
Notifications
You must be signed in to change notification settings - Fork 3.2k
MacOS M1 binary compilation and possibility to fine tune a model in C++ #11343
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
For your information, the macOS binaries in our github release package were built in a way like:
I think you don't need to use "--build_java --build_nodejs" And we haven't tried it for training. I know training x86_64 works. I haven't tried arm. I hope there won't be problems. |
Thanks! Now with this command line:
I get a different compilation error but still related to protobuf:
|
I haven't seen such error before. It looks like the compiler you are using isn't compatible with the protobuf version we are using. When the build just started, cmake should print out compiler information. Would you please post it here? Is it Apple Clang?_ |
Here the summary given by CMake:
And the compiler version that CMake uses:
|
And for your information, this is what we have now:
Yours is newer. I will try to see if I can get it one, otherwise it's difficult for me to work on. |
My version of Xcode is 13.3.1 and I'm on MacOS Monterey 12.3.1 if it may help you. |
@snnn Any update on your side on testing the compilation on a newer environment? |
Describe the bug
Issue to compile ONNX Runtime with training enabled.
Urgency
Would be perfect to get an answer before the end of the first week of May, means at most the 6th of May. As I have to propose a POC before that date.
System information
To Reproduce
Expected behavior
Properly compile ONNX Runtime
Additional context
The compilation error is:
I suspect to do not provide the proper parameters, but there is a lack of documentation on this part, the only thing I found is this page.
This bring me to another question, is training and fine tuning models from C++ with ONNX Runtime possible on MacOS M1 at least? If not it is useless for me to go deeper in this.
Thanks a lot in advance for any light you could provide on this issue.
The text was updated successfully, but these errors were encountered: