Closed MoisesHer closed 3 years ago
@szha It looks that there are something wrong with the latest MXNet master. Lots of GluonNLP test cases have failed. Would it be related to the graph optimization PR?
The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1432/fix_prepare_data_squad/index.html
@leezu @ZheyuYe To explain why we are changing from os.link
to shutil.copyfile
. We noticed some issues when running the experiment in the docker environment so we made the change. Also, SQuAD is not very large and it should be acceptable.
Should be unblocked temporarily by https://github.com/dmlc/gluon-nlp/pull/1433
@MoisesHer Would you try to merge the master?
The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1432/fix_prepare_data_squad/index.html
Merging #1432 (2ad77e0) into master (b9600b3) will decrease coverage by
0.20%
. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #1432 +/- ##
==========================================
- Coverage 85.51% 85.31% -0.21%
==========================================
Files 53 53
Lines 6987 6987
==========================================
- Hits 5975 5961 -14
- Misses 1012 1026 +14
Impacted Files | Coverage Δ | |
---|---|---|
src/gluonnlp/data/loading.py | 78.11% <0.00%> (-5.29%) |
:arrow_down: |
src/gluonnlp/data/tokenizers/yttm.py | 81.89% <0.00%> (-0.87%) |
:arrow_down: |
src/gluonnlp/data/tokenizers/subword_nmt.py | 79.43% <0.00%> (+0.93%) |
: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 b9600b3...2ad77e0. Read the comment docs.
The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1432/fix_prepare_data_squad/index.html
Description
This PR fixes the following issue found when getting SQuAD datasets on a MXNet docker container:
Docker version 19.03.8, build afacb8b7f0
Checklist
Essentials
cc @dmlc/gluon-nlp-team, @sxjscience