Closed barry-jin closed 4 years ago
Merging #1349 into master will increase coverage by
0.10%
. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #1349 +/- ##
==========================================
+ Coverage 81.65% 81.75% +0.10%
==========================================
Files 52 52
Lines 6862 6862
==========================================
+ Hits 5603 5610 +7
+ Misses 1259 1252 -7
Impacted Files | Coverage Δ | |
---|---|---|
src/gluonnlp/data/tokenizers/subword_nmt.py | 79.43% <0.00%> (+0.93%) |
:arrow_up: |
src/gluonnlp/data/loading.py | 83.39% <0.00%> (+2.26%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 203d84c...cab4c6d. Read the comment docs.
@szha Would you review/merge?
For me, I feel that we can have two artifacts: 1) the log of the unittest, 2) the complete log generated by the batch job. The advantage is that we may choose to download one of them to reduce the time for investigating the log.
Having two logs may be confusing. The problem is that the unittest can fail with cryptic errors if there is a problem in the setup phase of the batch job. People may not understand that they need to check a separate log file to figure out the error
@leezu The problem here is that the full log can have >400MB, which looks scary for most people. We may write a README to teach the user how to investigate batch-related problems.
The problem here is that the full log can have >400MB
That's due to misconfiguration.
Description
Add codecov report and upload batch job logs as build artifacts.
Checklist
Essentials
Changes
Comments
cc @dmlc/gluon-nlp-team