Historical Benchmark Performance Dashboard

[lapp0]

What behavior of the library made you think about the improvement?

Follow-up to #903 which implements ASV benchmarks comparing main to HEAD of PR.

With the infrastructure introduced in the above PR, we can create useful historical performance dashboards such as https://asv-runner.github.io/asv-collection/pandas/

How would you like it to behave?

• On merge, ASV benchmarks are run using main, and published to outlines-dev/outlines-benchmarks
• A github page is setup for outlines-benchmarks
• To ensure consistency over time, this requires a stable, dedicated machine which must have a guarantee of being idle during benchmark runs.