fix skip_learned_data to handle epoch/step based training

[senwu]

Description of the problems or issues

Is your pull request related to a problem? Please describe. The current implementation doesn't handle epoch based skip learned data.

Description of the proposed changes

The PR fixes the epoch based skip learned data training.

Test plan

Pass all and new tests.

Checklist

• [x] I have updated the documentation accordingly.
• [x] I have added tests to cover my changes.
• [x] All new and existing tests passed.
• [x] I have updated the CHANGELOG.rst accordingly.

[codecov[bot]]

Codecov Report

Merging #108 (e7d5d89) into master (bb7d359) will increase coverage by 1.34%. The diff coverage is 66.66%.

@@            Coverage Diff             @@
##           master     #108      +/-   ##
==========================================
+ Coverage   85.45%   86.80%   +1.34%     
==========================================
  Files          41       41              
  Lines        2022     2031       +9     
  Branches      437      438       +1     
==========================================
+ Hits         1728     1763      +35     
+ Misses        177      158      -19     
+ Partials      117      110       -7     

Flag	Coverage Δ
unittests	86.80% <66.66%> (+1.34%)	:arrow_up:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/emmental/learner.py	75.85% <66.66%> (+7.38%)	:arrow_up:
src/emmental/logging/logging_manager.py	86.44% <0.00%> (+0.84%)	:arrow_up:
src/emmental/logging/checkpointer.py	93.33% <0.00%> (+1.90%)	:arrow_up:
src/emmental/meta.py	76.69% <0.00%> (+1.94%)	:arrow_up: