|
a |
|
b/.gitmessage |
|
|
1 |
# Type: <subject> (<50-characters) |
|
|
2 |
|
|
|
3 |
# Body (Wrap it to 72-character) |
|
|
4 |
## What and Why you did: |
|
|
5 |
|
|
|
6 |
## The Impact of Change: |
|
|
7 |
* |
|
|
8 |
|
|
|
9 |
# Type should be one of the following: |
|
|
10 |
# * feat (new feature) |
|
|
11 |
# * fix (bug fix) |
|
|
12 |
# * docs (changes to documentation) |
|
|
13 |
# * style (formatting, missing semi colons, etc; no code change) |
|
|
14 |
# * refactor (refactoring production code) |
|
|
15 |
# * test (adding missing tests, refactoring tests; no production code change) |
|
|
16 |
# * chore (updating grunt tasks etc; no production code change) |
|
|
17 |
# |
|
|
18 |
# Example Commit Message |
|
|
19 |
# ------------------------------------------------|----------------------| |
|
|
20 |
# feat: Add FancyNormLayer to MyDeepNet |
|
|
21 |
# what: Implement a normalization layer presented in [http:~]. |
|
|
22 |
# why: This layer would work better than the current norm layers. |
|
|
23 |
# * Potentially improves classfication accuracy by normalizing input to |
|
|
24 |
# each convolution layers. |
|
|
25 |
# * The paper also reports that the layer would contribute to faster |
|
|
26 |
# convergence with proper hyper parameter of the norm layers. |
|
|
27 |
# * As side effect, the performance may get unstable if mini batch size |
|
|
28 |
# is small. |
|
|
29 |
# ------------------------------------------------|----------------------| |
|
|
30 |
# 72-character wrapped longer description. |
|
|
31 |
# This template is based on: |
|
|
32 |
# https://thoughtbot.com/blog/better-commit-messages-with-a-gitmessage-template |
|
|
33 |
# https://gist.github.com/Linell/bd8100c4e04348c7966d |
|
|
34 |
# |