Downloads · 30 days
4
1% of all-time downloads
LandmarkNER
constantinSch/LandmarkNER
LandmarkNER is a token classification model from constantinSch. Use it when you need labels on individual words, such as names. It is set up for spacy.
German Named Entity Recognition model for recognizing Bavarian landmarks. Fine-tuned "bert-base-german-cased" with 6450 annotated sentences of which 1467 contained landmarks, from subtitles of videos from Bayerischer…
All-time downloads
460
Public
Repo size
843 MB
Likes
1
Public
Repo makeup
Click a slice to open those files.
Other438 MB · 52%
At a glance
- Task
- Token Classification
- Library
- spacy
- Access
- Public
- Created
- Jul 25, 2022
- Updated
- Jul 27, 2022
- SHA
- aa29f430
Model card
From the Hugging Face model README
Introduction
German Named Entity Recognition model for recognizing Bavarian landmarks. Fine-tuned "bert-base-german-cased" with 6450 annotated sentences of which 1467 contained landmarks, from subtitles of videos from Bayerischer Rundfunk.
| Feature | Description |
|---|---|
| Name | de_pipeline |
| Version | 0.1.0 |
| spaCy | >=3.3.0,<3.4.0 |
| Default Pipeline | transformer, ner |
| Components | transformer, ner |
| Vectors | 0 keys, 0 unique vectors (0 dimensions) |
| Sources | n/a |
| License | n/a |
| Author | Constantin Förster |
Label Scheme
<details> <summary>View label scheme (1 labels for 1 components)</summary>| Component | Labels |
|---|---|
ner | LM |
Accuracy
| Type | Score |
|---|---|
ENTS_F | 85.39 |
ENTS_P | 83.79 |
ENTS_R | 87.05 |
TRANSFORMER_LOSS | 4216.96 |
NER_LOSS | 78511.31 |
- Task
- Token Classification
- Library
- spacy
- Languages
- de
- Created
- Jul 25, 2022
- Updated
- Jul 27, 2022