SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2010.11934 · 2020

mT5: A massively multilingual pre-trained text-to-text transformer

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 13 functions out of this paper's own repositories and ran 10 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
google-research/multilingual-t5 canonical 5 of 5
KoshiroSato/Flask_NLP_App pwc_unofficial 4 of 7
MorenoLaQuatra/bart-it pwc_unofficial 1 of 1
FunctionStatusWhere it lives
bert_predict Ran KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/bert_model.py
code served (permissive licence) · get_code("ab4db4a59f94ef40")
mt5_post_processing Ran KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/mt5_model.py
code served (permissive licence) · get_code("0d9ef0d0dfe26859")
normalize_mlqa Ran google-research/multilingual-t5/multilingual_t5/evaluation/metrics.py
code served (permissive licence) · get_code("4ad4262f32bec052")
predict_with_post_processing Ran KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/bert_model.py
code served (permissive licence) · get_code("0d4436476da0cc30")
preprocess_logits_for_metrics Ran MorenoLaQuatra/bart-it/summarization/finetune_summarization.py
code served (permissive licence) · get_code("2641f706223d1e5c")
read_txt Ran KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/utils.py
code served (permissive licence) · get_code("e445d0574fe31243")
span_f1 Ran google-research/multilingual-t5/multilingual_t5/evaluation/metrics.py
code served (permissive licence) · get_code("3a70b94fccf1c165")
wikiann Ran google-research/multilingual-t5/multilingual_t5/preprocessors.py
code served (permissive licence) · get_code("6a5b181d28b784b6")
xnli_map_hypothesis_premise Ran google-research/multilingual-t5/multilingual_t5/preprocessors.py
code served (permissive licence) · get_code("f04d1c1706891df6")
xquad Ran google-research/multilingual-t5/multilingual_t5/preprocessors.py
code served (permissive licence) · get_code("64c4f17ea92b71d7")
app_logger Not yet run KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/logger.py
code served (permissive licence) · get_code("d10b1ac136fbfec1")
model_io_logger Not yet run KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/logger.py
code served (permissive licence) · get_code("baf8e460eed595db")
read_yaml Not yet run KoshiroSato/Flask_NLP_App/flask/flaskapp/lib/utils.py
code served (permissive licence) · get_code("c9737e28fdfdcfe9")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

The recent "Text-to-Text Transfer Transformer" (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 that was pre-trained on a new Common Crawl-based dataset covering 101 languages. We detail the design and modified training of mT5 and demonstrate its state-of-the-art performance on many multilingual benchmarks. We also describe a simple technique to prevent "accidental translation" in the zero-shot setting, where a generative model chooses to (partially) translate its prediction into the wrong language. All of the code and model checkpoints used in this work are publicly available.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2010.11934")
get_code_for_paper("2010.11934")
have("2010.11934")

Connect an agent — have() is free.