Add quantized training (CPU part)

[shiyu1994]

This is to add the quantized training for CPU, which speed up the training. (#5606)

[shiyu1994]

It is weird that the test_monotone_penalty_max fails in this commit [6a76fde], which doesn't seem to have any changes that will cause the failure. This should call for further attention to check if the test case fails randomly in some cases. Maybe we should simply make this test case more deterministic by setting the number of threads to be 1 or enable the deterministic flag.

(6a76fde)
https://dev.azure.com/lightgbm-ci/lightgbm-ci/_build/results?buildId=14615&view=logs&j=c28dceab-947a-5848-c21f-eef3695e5f11&t=fa158246-17e2-53d4-5936-86070edbaacf

[shiyu1994]

@guolinke This is ready for review.

[shiyu1994]

@guolinke This is ready for review.

Sorry I think one more test case for distributed quantized training should be added in test_dask.py. Please pending the review until that is ready.

[shiyu1994]

@guolinke This is ready for review.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed.
To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues
including a reference to this.