Add facilities for unit and functional tests

[danmcp]

There was already 1 unit test in place. The existing functional tests are being moved to scripts for consistency with other projects.

One new functional test has been added to test the mt_bench_branch generator as a starting example

I've included basic tests for mt_bench and mmlu. Which already found two issues:

• mmlu_branch was using self.tasks instead of the number of results calculating the wrong average
• Not specifying max_workers for mt_bench resulted in an error when it should have defaulted

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. @danmcp please rebase it. https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[nathan-weinberg]

Looked over all these changes and honestly everything LGTM - thank you for this @danmcp!

Two nits - one is in the comment, the other would be it'd be nice to add a badge for this new workflow similar to what is done for the CLI here: https://github.com/instructlab/instructlab/blob/main/README.md?plain=1#L4

[danmcp]

the other would be it'd be nice to add a badge for this new workflow similar to what is done for the CLI here: https://github.com/instructlab/instructlab/blob/main/README.md?plain=1#L4

Done!