Support multiple device training and inference

[senwu]

Description of the problems or issues

Support training and evaluate model using mix devices

Description of the proposed changes

Specify module device using module_device.

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 #91 (38935e1) into master (0d69bb3) will increase coverage by 0.45%. The diff coverage is 75.60%.

@@            Coverage Diff             @@
##           master      #91      +/-   ##
==========================================
+ Coverage   87.45%   87.90%   +0.45%     
==========================================
  Files          38       38              
  Lines        1753     1777      +24     
  Branches      379      385       +6     
==========================================
+ Hits         1533     1562      +29     
+ Misses        116      108       -8     
- Partials      104      107       +3     

Flag	Coverage Δ
unittests	87.90% <75.60%> (+0.45%)	:arrow_up:

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

Impacted Files	Coverage Δ
src/emmental/task.py	83.33% <50.00%> (-12.13%)	:arrow_down:
src/emmental/model.py	81.02% <81.48%> (+0.96%)	:arrow_up:
src/emmental/utils/utils.py	92.62% <83.33%> (+7.49%)	:arrow_up: