Evaluation on Benchmarks

#32
by antogrk - opened

Thank you for the amazing work and for sharing this model!

Would it be possible to provide any specific scripts that you used to evaluate the model on the different benchmarks (e.g. MMMU, MMBench, ScienceQA etc.)?

Microsoft org

@antogrk Thank you for your interest in Phi-4-multimodal.
We do not open source our evaluation platform you can read more about our methodology in the model card.
The model is available publicly, so we also want to see the independent evaluation from community, for example this one.

Sign up or log in to comment