dmlc / gluon-nlp

NLP made easy

https://nlp.gluon.ai/

Apache License 2.0

2.55k stars 538 forks source link

[CONVERTER] fix fairseq conversion scripts #1466

Closed szha closed 3 years ago

szha commented 3 years ago

Description

fix fairseq conversion scripts with current upstream

Checklist

Essentials

[x] PR's title starts with a category (e.g. [BUGFIX], [MODEL], [TUTORIAL], [FEATURE], [DOC], etc)
[x] Changes are complete (i.e. I finished coding on this PR)
[x] All changes have test coverage
[x] Code is well-documented

Changes

[x] fix fairseq conversion scripts with current upstream

Comments

the config object was moved from the HubInterface.args to HubInterface.cfg.model.

cc @dmlc/gluon-nlp-team

sxjscience commented 3 years ago

LGTM. I think we can add tests of these conversion scripts later.

szha commented 3 years ago

Thanks for the quick review. I looked at the conversion scripts and thought that the most meaningful test for them would probably just be the successful conversion of parameters, which will be verified once we have scheduled jobs for running them regularly and for reporting failure

codecov[bot] commented 3 years ago

Codecov Report

Merging #1466 (2c03c15) into master (12f6da2) will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1466   +/-   ##
=======================================
  Coverage   85.84%   85.84%           
=======================================
  Files          52       52           
  Lines        6907     6907           
=======================================
  Hits         5929     5929           
  Misses        978      978

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 12f6da2...2c03c15. Read the comment docs.

sxjscience commented 3 years ago

The GPU test has failed. It might be related to some errors in MXNet.

github-actions[bot] commented 3 years ago

The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1466/fix_fairseq_convert/index.html