feat: Support for handling Benchmarks V2

[jstlaurent]

Changelogs

• Make BenchmarkSpecification into an abstract class, with BenchmarkV1Specification and BenchmarkV2Specification implementations. The abstract class defines a series of abstract methods and properties
  for the subclasses to implement.
• Implemented a new type of split for Benchmark V2, using the Roaring Bitmap format to efficiently store extremely large sets of indicices.
• Improved the type definition in the Pydantic models, to reflect the post-validation types. Where we are flexible on the input type, validators with mode="before" ensure that we convert the input to a
  single internal representation.
• Added upload/dowload methods to the Hub client, to support interacting with the Hub storage.

---

Checklist:

• Was this PR discussed in an issue? It is recommended to first discuss a new feature into a GitHub issue before opening a PR.
• Add tests to cover the fixed bug(s) or the newly introduced feature(s) (if appropriate).
• Update the API documentation if a new function is added, or an existing one is deleted.
• Write concise and explanatory changelogs above.
• If possible, assign one of the following labels to the PR: feature, fix, chore, documentation or test (or ask a maintainer to do it for you).

---

The core changes are ready for review. I need to add some tests and ensure the doc is updated as well.

[cwognum]

Thanks @jstlaurent and @mercuryseries ! This is looking good! 🚀

[mercuryseries]

Looks good to me, @jstlaurent! I added some minor comments here and there.

try:
    while True:
        print("Thank you! (Until Julien is tired)")
except KeyboardInterrupt:
    print("Gratitude interrupted, but it’s endless in spirit! :)")