Downloads · 30 days
6
10% of all-time downloads
marialcasimiro/tatoeba-opus-2021-02-22-fra-eng
tatoeba-opus-2021-02-22-fra-eng is a translation model from marialcasimiro. Use it when you need text moved from one language to another. It is set up for transformers. The card lists the license as apache-2.0.
source language name: French target language name: English OPUS readme: README.md
Downloads · 30 days
6
10% of all-time downloads
All-time downloads
63
Public
Parameters
111M
1.3 GB on disk
Likes
0
Public
Click a slice to open those files.
.bin222 MB · 50%
From the Hugging Face model README
source language name: French
target language name: English
OPUS readme: README.md
model: transformer-align
source language code: fr
target language code: en
dataset: opus
release date: 2021-02-22
pre-processing: normalization + SentencePiece (spm32k,spm32k)
download original weights: opus-2021-02-22.zip
Training data:
Validation data:
Test data:
test set translations file: test.txt
test set scores file: eval.txt
BLEU-scores |Test set|score| |---|---| |Tatoeba-test.fra-eng|57.8| |newsdiscusstest2015-enfr.fra-eng|39.7| |newstest2014-fren.fra-eng|38.4| |newsdiscussdev2015-enfr.fra-eng|34.4| |newstest2013.fra-eng|34.0| |newstest2012.fra-eng|33.2| |newstest2011.fra-eng|33.1| |newstest2010.fra-eng|32.7| |newssyscomb2009.fra-eng|31.1| |newstest2009.fra-eng|30.5| |news-test2008.fra-eng|26.5|
chr-F-scores |Test set|score| |---|---| |Tatoeba-test.fra-eng|0.723| |newstest2014-fren.fra-eng|0.636| |newsdiscusstest2015-enfr.fra-eng|0.621| |newstest2011.fra-eng|0.598| |newstest2010.fra-eng|0.593| |newstest2012.fra-eng|0.593| |newstest2013.fra-eng|0.592| |newsdiscussdev2015-enfr.fra-eng|0.587| |newssyscomb2009.fra-eng|0.575| |newstest2009.fra-eng|0.572| |news-test2008.fra-eng|0.544|