Closed xcgoner closed 4 years ago
Job PR-1277/1 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1277/1/index.html
Merging #1277 into master will increase coverage by
0.27%
. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #1277 +/- ##
==========================================
+ Coverage 87.45% 87.72% +0.27%
==========================================
Files 81 81
Lines 7365 7365
==========================================
+ Hits 6441 6461 +20
+ Misses 924 904 -20
Impacted Files | Coverage Δ | |
---|---|---|
src/gluonnlp/data/word_embedding_evaluation.py | 96.93% <0.00%> (+7.66%) |
:arrow_up: |
@sxjscience @szha I've updated the results reported in the webpage in this pr. Although the results of language models look fine to me, the cached language models seem to have a significant performance regression. The updated logs could be found in https://github.com/dmlc/web-data/pull/259
Description
Resolve the issue https://github.com/dmlc/gluon-nlp/issues/1253 Currently, the average trigger checks “val_L > min(valid_losses[-n:])”. In this patch, I change the implementation into the version published in ICLR, [1], the algorithms is actually “val_L > min(valid_losses[:-n])”, which is also used in Salesforce’s source code: [2]
References
[1] https://openreview.net/pdf?id=SyyGPP0TZ [2] https://github.com/salesforce/awd-lstm-lm/blob/32fcb42562aeb5c7e6c9dec3f2a3baaaf68a5cb5/main.py#L275
cc @dmlc/gluon-nlp-team