naetimus / flair-lms

Language Models for Zalando's flair library

flair-lms

This repository is part of the NLP research with flair, a state-of-the-art NLP framework from Zalando Research.

This repository will include various language models (forward and backward) that can be used with flair. It will be updated frequently. So please star or watch this repository 😅

Changelog

January 2020: Move repository to the new FlairNLP group on GitHub.

September 2019: New Multilingual Flair Embeddings trained on JW300 corpus are released.

September 2019: All Flair Embeddings that are now officially available in Flair >= 0.4.3 are listed.

Parameters

All Flair Embeddings are trained with a hidden_size of 2048 and nlayers of 1.

Flair Embeddings

Language model	# Tokens	Forward ppl.	Backward ppl.	Flair Embeddings alias
Arabic	736M	3.39	3.45	ar-forward and ar-backward
Bulgarian (fast)	66M	2.48	2.51	bg-forward-fast and bg-backward-fast
Bulgarian	111M	2.46	2.47	bg-forward and bg-backward
Czech (v0)	778M	3.44	3.48	cs-v0-forward and cs-v0-backward
Czech	442M	2.89	2.90	cs-forward and cs-backward
Danish	325M	2.62	2.68	da-forward and da-backward
Basque (v0)	37M	2.56	2.58	eu-v0-forward and eu-v0-backward
Basque (v1)	37M	2.64	2.31	eu-v1-forward and eu-v1-backward
Basque	57M	2.90	2.83	eu-forward and eu-backward
Persian	146M	3.68	3.66	fa-forward and fa-backward
Finnish	427M	2.63	2.65	fi-forward and fi-backward
Hebrew	502M	3.84	3.87	he-forward and he-backward
Hindi	28M	2.87	2.86	hi-forward and hi-backward
Croatian	625M	3.13	3.20	hr-forward and hr-backward
Indonesian	174M	2.80	2.74	id-forward and id-backward
Italian	1,5B	2.62	2.63	it-forward and it-backward
Dutch (v0)	897M	2.78	2.77	nl-v0-forward and nl-v0-backward
Dutch	1,2B	2.43	2.55	nl-forward and nl-backward
Norwegian	156M	3.01	3.01	no-forward and no-backward
Polish	1,4B	2.95	2.84	pl-opus-forward and pl-opus-backward
Slovenian (v0)	314M	3.28	3.34	sl-v0-forward and sl-v0-backward
Slovenian	419M	2.88	2.91	sl-forward and sl-backward
Swedish (v0)	545M	2.29	2.27	sv-v0-forward and sv-v0-backward
Swedish	671M	6.82 (?)	2.25	sv-forward and sv-backward
Tamil	18M	2.23	4509 (!)	ta-forward and ta-backward

Multilingual Flair Embeddings

Multilingual Flair Embeddings were trained on the recently released JW300 corpus. Thanks to half precision support in Flair, both forward and backward Embeddings were trained for 5 epochs for over 10 days. The training corpus has 2,025,826,977 token.

Language model	# Tokens	Forward ppl.	Backward ppl.	Flair Embeddings alias
JW300	2B	3.25	3.37	multi-forward and multi-backward

It can be loaded with:

from flair.embeddings import FlairEmbeddings

jw_forward = FlairEmbeddings("multi-forward")
jw_backward = FlairEmbeddings("multi-backward")

A detailed evaluation on various PoS tagging tasks can be found in this repository.

We would like to thank Željko Agić for providing us access to the corpus (before it was officially released)!