Downloads · 30 days
10
1% of all-time downloads
flair-historic-ner-lft
dbmdz/flair-historic-ner-lft
flair-historic-ner-lft is a token classification model from dbmdz. Use it when you need labels on individual words, such as names. It is set up for flair. The card lists the license as mit.
Based on our paper we release a new model trained on the LFT dataset.
All-time downloads
1.5K
Public
Repo size
888 MB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.bin444 MB · 100%
At a glance
- Task
- Token Classification
- Library
- flair
- License
- mit
- Access
- Public
- Created
- Mar 2, 2022
- Updated
- Dec 11, 2020
- SHA
- 2fd76829
Model card
From the Hugging Face model README
Towards Robust Named Entity Recognition for Historic German
Based on our paper we release a new model trained on the LFT dataset.
Note: We use BPEmbeddings instead of the combination of Wikipedia, Common Crawl and character embeddings (as used in the paper), so save space and training/inferencing time.
Results
| Dataset \ Run | Run 1 | Run 2 | Run 3† | Avg. |
|---|---|---|---|---|
| Development | 76.32 | 76.13 | 76.36 | 76.27 |
| Test | 77.07 | 77.35 | 77.20 | 77.21 |
Paper reported an averaged F1-score of 77.51.
† denotes that this model is selected for upload.
- Task
- Token Classification
- Library
- flair
- License
- mit
- Languages
- de
- Created
- Mar 2, 2022
- Updated
- Dec 11, 2020