[Tune] Inclusion of SOTA HPO algorithms

[BradKML]

Description

Was browsing GitHub to see which library has the widest amount of HPO algorithms besides Ray (which relies on other libraries), and these 3 popped up:

• https://orion.readthedocs.io/en/stable/user/algorithms.html
• https://syne-tune.readthedocs.io/en/latest/getting_started.html
• https://oss-vizier.readthedocs.io/en/latest/guides/user/supported_algorithms.html

Use case

There has to be faster algorithms than just BOHB or ASHA that is worth trying, for data that is too large, and training time that is too long for each iteration. Thinking what other algorithms can be done to make this easier. How can they be compatible with Ray?