Hugging Face's logo

chcaa
/
da_dacy_tiny

Token Classification
spaCy
Danish
dacy
danish
pos tagging
morphological analysis
dependency parsing
named entity recognition
Eval Results (legacy)
Model card Files
xet
 Community
3
johanamayer commited on
Commit
f361251
·
verified ·
1 Parent(s): ce0ab11

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +19 -8
README.md CHANGED
@@ -122,17 +122,27 @@ datasets:
122
  metrics:
123
  - accuracy
124
  ---
 
 
 
 
 
 
 
 
 
 
125
  | Feature | Description |
126
  | --- | --- |
127
- | **Name** | `da_dacy_tiny_dacy_large` |
128
  | **Version** | `0.3.0` |
129
  | **spaCy** | `>=3.8.14,<3.9.0` |
130
  | **Default Pipeline** | `tok2vec`, `lemmatizer`, `tagger`, `morphologizer`, `parser`, `ner` |
131
  | **Components** | `tok2vec`, `lemmatizer`, `tagger`, `morphologizer`, `parser`, `ner` |
132
  | **Vectors** | 500000 keys, 500000 unique vectors (300 dimensions) |
133
- | **Sources** | n/a |
134
- | **License** | n/a |
135
- | **Author** | [n/a]() |
136
 
137
  ### Label Scheme
138
 
@@ -156,9 +166,6 @@ metrics:
156
  | `LEMMA_ACC` | 95.40 |
157
  | `TAG_ACC` | 96.55 |
158
  | `POS_ACC` | 96.55 |
159
- | `TAG_MICRO_P` | 0.00 |
160
- | `TAG_MICRO_R` | 0.00 |
161
- | `TAG_MICRO_F` | 0.00 |
162
  | `MORPH_ACC` | 96.21 |
163
  | `DEP_UAS` | 84.31 |
164
  | `DEP_LAS` | 80.25 |
@@ -173,4 +180,8 @@ metrics:
173
  | `TAGGER_LOSS` | 68699.56 |
174
  | `MORPHOLOGIZER_LOSS` | 103458.12 |
175
  | `PARSER_LOSS` | 1188327.34 |
176
- | `NER_LOSS` | 49397.07 |
 
 
 
 
 
122
  metrics:
123
  - accuracy
124
  ---
125
+ <a href="https://github.com/centre-for-humanities-computing/Dacy"><img src="https://centre-for-humanities-computing.github.io/DaCy/_static/icon.png" width="175" height="175" align="right" /></a>
126
+
127
+ # DaCy tiny
128
+
129
+ DaCy is a Danish language processing framework with state-of-the-art pipelines as well as functionality for analysing Danish pipelines.
130
+ This model is the **"tiny" Dacy pipeline**, trained on da_news_core_lg vectors for Danish linguistic
131
+ annotation and downstream NLP tasks.
132
+ To read more check out the [DaCy repository](https://github.com/centre-for-humanities-computing/DaCy) for material on how to use DaCy and reproduce the results.
133
+ DaCy also contains guides on usage of the package as well as behavioural test for biases and robustness of Danish NLP pipelines.
134
+
135
  | Feature | Description |
136
  | --- | --- |
137
+ | **Name** | `dacy_tiny` |
138
  | **Version** | `0.3.0` |
139
  | **spaCy** | `>=3.8.14,<3.9.0` |
140
  | **Default Pipeline** | `tok2vec`, `lemmatizer`, `tagger`, `morphologizer`, `parser`, `ner` |
141
  | **Components** | `tok2vec`, `lemmatizer`, `tagger`, `morphologizer`, `parser`, `ner` |
142
  | **Vectors** | 500000 keys, 500000 unique vectors (300 dimensions) |
143
+ | **Sources** | [UD Danish DDT v2.18](https://github.com/UniversalDependencies/UD_Danish-DDT) (Johannsen, Anders; Martínez Alonso, Héctor; Plank, Barbara)<br>[DaNE](https://huggingface.co/datasets/alexandrainst/dane) (Rasmus Hvingelby, Amalie B. Pauli, Maria Barrett, Christina Rosted, Lasse M. Lidegaard, Anders Søgaard)<br>[spacy/da_core_news_lg](https://huggingface.co/spacy/da_core_news_lg) (Explosion) |
144
+ | **License** | `Apache-2.0` |
145
+ | **Author** | [Kenneth Enevoldsen, Johana Mayerova, Mikkel Krøjen Svendsen](https://chc.au.dk/)|
146
 
147
  ### Label Scheme
148
 
 
166
  | `LEMMA_ACC` | 95.40 |
167
  | `TAG_ACC` | 96.55 |
168
  | `POS_ACC` | 96.55 |
 
 
 
169
  | `MORPH_ACC` | 96.21 |
170
  | `DEP_UAS` | 84.31 |
171
  | `DEP_LAS` | 80.25 |
 
180
  | `TAGGER_LOSS` | 68699.56 |
181
  | `MORPHOLOGIZER_LOSS` | 103458.12 |
182
  | `PARSER_LOSS` | 1188327.34 |
183
+ | `NER_LOSS` | 49397.07 |
184
+
185
+ ### Training
186
+
187
+ This model was trained using [spaCy](https://spacy.io)