Go to file
2021-02-08 21:42:08 +01:00
data-bin fairseq v1 2021-02-07 01:01:44 +01:00
dev-0 Update repo 2020-10-26 20:26:04 +01:00
src fairseq v1 2021-02-07 01:01:44 +01:00
test-A next test 2021-02-08 21:42:08 +01:00
train Update repo 2020-10-26 20:26:04 +01:00
.gitignore Update repo 2020-10-26 20:26:04 +01:00
command-preprocess fairseq v1 2021-02-07 01:01:44 +01:00
config.txt Update repo 2020-10-26 20:26:04 +01:00
gru_attention.ipynb GRU very simple 2021-02-08 14:36:14 +01:00
README.md Update repo 2020-10-26 20:26:04 +01:00
requirements.txt fairseq v1 2021-02-07 01:01:44 +01:00
train_command fairseq v1 2021-02-07 01:01:44 +01:00

wmt-2020-pl-en

Translate from Polish to English.

This is a challenge created from http://www.statmt.org/wmt20/translation-task.html . Train set is created from europarl wmt pl-en training data. Dev and test set are created from wmt pl-en development data.

Directory structure

  • README.md — this file
  • config.txt — configuration file
  • train/ — directory with training data
  • train/train.tsv — sample parallel corpus (Finnish text in the first column, Māori text in the second one)
  • dev-0/ — directory with dev (test) data
  • dev-0/in.tsv — Finnish input text for the dev set
  • dev-0/expected.tsv — Māori reference translation for the dev set
  • test-A — directory with test data
  • test-A/in.tsv — Finnish input data for the test set
  • test-A/expected.tsv — Māori reference translation for the test set