Tool for mining/alignment parallel texts
Project description
hadal
hadal /ˈheɪdəl/
is a tool for parallel sentence mining with pretrained models.
🚧🚧🚧The project is under active development. The changes in the API can be breaking. 🚧🚧🚧
Installation
See INSTALL.md for detailed installation instructions.
Implemented methods
Method | Alignment type |
---|---|
margin-based | one-to-one |
soon... |
Quickstart
See demo file demo.py for more details.
$ python demo.py
# score, source_sentence, target_sentence
[
(1.5549, "I think I like wine now.", "Je pense que j'aime le vin maintenant."),
(1.5079, "She eats one apple every day.", "Elle mange une pomme chaque jour."),
(1.4353, "They serve pizza dogs in the cafeteria.", "Ils vendent des hot-dogs à la cafétéria."),
(0.4112, "Empty sentence.", "Ce jeu se joue sur le vaisseau spatial."),
]
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
hadal-0.0.1.tar.gz
(36.3 kB
view hashes)
Built Distribution
hadal-0.0.1-py3-none-any.whl
(39.8 kB
view hashes)