feat: Perform dataset mixing via sampling probabilities in data config

[dushyantbehl]

Description of the change

Related issue number

How to verify the PR

Was the PR tested

• I have added >=1 unit test(s) for every new method I have added.
• I have ensured all unit tests pass

[github-actions]

Thanks for making a pull request! 😃
One of the maintainers will review and advise on the next steps.

[dushyantbehl]

@ashokponkumar @willmj @Abhishek-TAMU the data mixing feature is ready for review now.

[ashokponkumar]

@dushyantbehl How do we plan to support validation split and training split?

[Abhishek-TAMU]

Thanks @dushyantbehl, for the PR. The code to add the feature of sampling from multiple datasets looks good to me. The unit tests also look good. Just minor suggestions.

[willmj]

Made an image for testing: https://v3.travis.ibm.com/github/ai-foundation/sft-trainer-image/builds/31001438

[dushyantbehl]

@dushyantbehl How do we plan to support validation split and training split?

@ashokponkumar as discussed offline this feature will be left to subsequent patches we apply so current patch is independent of this feature thanks.

[willmj]

LGTM! Thanks Dushyant

[Abhishek-TAMU]

Changes look good to me!