fix iterator when loading from checkpoint

[000Justin000]

Before submitting

• [no] Was this discussed/approved via a Github issue? (no need for typos, doc improvements)
• [yes] Did you read the contributor guideline?
• [yes] Did you make sure to update the docs?
• [yes] Did you write any new necessary tests?

What does this PR do?

Assuming an epoch consist of N batches. When we load a checkpoint from mid epoch, there is a number indicate the number of iterations n has been passed in the epoch. When restarting the training job, after the batches are loaded, n batches are removed from the beginning of the iterator by FrozenBatchSampler. Then the CounteringIterator would add n iteration count in front of the batch, so that the batch count in the epoch is correct, i.e. N-n+n = N.
However, when skip_reminder_batch is set to True, the CounteringIterator would truncate, incorrectly, the batch count from N batches to N-n, causing the epoch to have less number of total batches. As a result, you would see a jump in the fblearner ui on the progress --- not because the iteration counter increase, but rather because the total number of batches decrease.
This diffs would fix the problem above.

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

Did you have fun?

Make sure you had fun coding 🙃