diff --git a/scripts/utils.py b/scripts/utils.py index 0ec872500e936fd8bb4abdd32654bd8e7f933625..f1d03feb973a69f90662574b0803eccd37b36dba 100644 --- a/scripts/utils.py +++ b/scripts/utils.py @@ -17,9 +17,15 @@ LANG2TRANSFORMER = { "it": "dbmdz/bert-base-italian-cased", "ru": "blinoff/roberta-base-russian-v0", "sv": "KB/bert-base-swedish-cased", + "uk": "/tmp/lustre_shared/mklimasz/transformers/wikibert-base-uk-cased/", + "ta": "/tmp/lustre_shared/mklimasz/transformers/wikibert-base-ta-cased/", + "sk": "/tmp/lustre_shared/mklimasz/transformers/wikibert-base-sk-cased/", + "lt": "/tmp/lustre_shared/mklimasz/transformers/wikibert-base-lt-cased/", + "cs": "/tmp/lustre_shared/mklimasz/transformers/wikibert-base-cs-cased/", + "et": "/tmp/lustre_shared/mklimasz/transformers/etwiki-bert/", # "uk": http://dl.turkunlp.org/wikibert/wikibert-base-uk-cased/ # "ta": http://dl.turkunlp.org/wikibert/wikibert-base-ta-cased/ - # "sk": http://dl.turkunlp.org/wikibert/wikibert-base-sl-cased/ + # "sk": http://dl.turkunlp.org/wikibert/wikibert-base-sk-cased/ # "lt": http://dl.turkunlp.org/wikibert/wikibert-base-lt-cased/ # "lv": http://dl.turkunlp.org/wikibert/wikibert-base-lv-cased/ # "et": http://dl.turkunlp.org/estonian-bert/etwiki-bert/pytorch/