Request for precomputed inference results

[Bouhlal-code]

Hi,

I am currently trying to reproduce the DI-Bench results.

The dataset and repository instances work fine, but I am running into issues during the dependency inference step due to API rate limits (OpenAI), which prevents generating prediction files.

I wanted to ask:
Do you provide any precomputed inference results (prediction files) for the dataset (especially the large Python set), so that I can directly run the evaluation step?

This would help verify that my setup is correct before running the full inference.

Thanks a lot for your work!