We've created a fork with OpenAI API support. This allows for faster portability between various accelerators (AMD, Intel, etc...).
Additionally, we modified the evaluation code to also support CPU during the evaluation phases.
Are you looking at actively maintaining this, and would you be interested in a PR with changes?
We've created a fork with OpenAI API support. This allows for faster portability between various accelerators (AMD, Intel, etc...).
Additionally, we modified the evaluation code to also support CPU during the evaluation phases.
Are you looking at actively maintaining this, and would you be interested in a PR with changes?