Closed sxjscience closed 3 years ago
The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1433/fix_amp/index.html
The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1433/fix_amp/index.html
Merging #1433 (1f581e0) into master (84391ef) will decrease coverage by
0.25%
. The diff coverage is0.00%
.
@@ Coverage Diff @@
## master #1433 +/- ##
==========================================
- Coverage 85.55% 85.30% -0.26%
==========================================
Files 53 53
Lines 6987 6987
==========================================
- Hits 5978 5960 -18
- Misses 1009 1027 +18
Impacted Files | Coverage Δ | |
---|---|---|
src/gluonnlp/attention_cell.py | 73.72% <0.00%> (-6.67%) |
:arrow_down: |
src/gluonnlp/utils/parameter.py | 92.30% <0.00%> (-1.93%) |
:arrow_down: |
src/gluonnlp/data/tokenizers/subword_nmt.py | 78.50% <0.00%> (-0.94%) |
:arrow_down: |
src/gluonnlp/data/tokenizers/yttm.py | 82.75% <0.00%> (+0.86%) |
: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 84391ef...1f581e0. Read the comment docs.
The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1433/fix_amp/index.html
Description
Tiny fix of the amp for log_maskedsoftmax
Checklist
Essentials
cc @dmlc/gluon-nlp-team