Releases: aymara/lima-models
Releases · aymara/lima-models
Segmentation and Tagging - multi-objective
This is a partial release of Deeplima models, including segmentation and tagging models, as well as a few lemmatization models.
These models have been trained with a multi-objective optimization strategy. The best ones for tokens, upos and lemmas have been selected.
Tagging models
v0.1.6-beta-morph Update README.md
Embeddings for English (1/10 of C4)
v0.1.6-beta-en-c4medium Update README.md
Tokenization models
Tokenization only (without sentence segmentation) models trained on all trainable corpora from UD 2.8 collection (67 languages), 8 models for each corpus.
Deeplima commit: aymara/lima@40ab9ad
LSTM hidden state: 8
v0.1.5
Models for 60+ languages
v0.1.4-beta Fix path to copyright
MorphoSyntax models for English and French
Starting from this version models are packaged per language.
v0.1.2-beta
Models are installed to /usr/.
Tokenization models for English
v0.1.1-beta English tokenization models