Hugging Face's logo

hmbert
/
flair-icdar-nl

Token Classification
Flair
PyTorch
Dutch
sequence-tagger-model
Model card Files
xet
 Community
1
stefan-it commited on
Commit
1d9754e
·
1 Parent(s): 806e269

readme: add initial version of model card

Browse files

Hi,

this PR adds the initial version of the model card.

Files changed (1) hide show
  1. README.md +74 -0
README.md ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: nl
3
+ license: mit
4
+ tags:
5
+ - flair
6
+ - token-classification
7
+ - sequence-tagger-model
8
+ base_model: dbmdz/bert-base-historic-multilingual-cased
9
+ widget:
10
+ - text: Professoren der Geneeskun dige Faculteit te Groningen alsook van de HH , Doctoren
11
+ en Chirurgijns van Groningen , Friesland , Noordholland , Overijssel , Gelderland
12
+ , Drenthe , in welke Provinciën dit Elixir als Medicament voor Mond en Tanden
13
+ reeds jaren bakend is .
14
+ ---
15
+
16
+ # Fine-tuned Flair Model on Dutch ICDAR-Europeana NER Dataset
17
+
18
+ This Flair model was fine-tuned on the
19
+ [Dutch ICDAR-Europeana](https://github.com/stefan-it/historic-domain-adaptation-icdar)
20
+ NER Dataset using hmBERT as backbone LM.
21
+
22
+ The ICDAR-Europeana NER Dataset is a preprocessed variant of the
23
+ [Europeana NER Corpora](https://github.com/EuropeanaNewspapers/ner-corpora) for Dutch and French.
24
+
25
+ The following NEs were annotated: `PER`, `LOC` and `ORG`.
26
+
27
+ # Results
28
+
29
+ We performed a hyper-parameter search over the following parameters with 5 different seeds per configuration:
30
+
31
+ * Batch Sizes: `[8, 4]`
32
+ * Learning Rates: `[3e-05, 5e-05]`
33
+
34
+ And report micro F1-score on development set:
35
+
36
+ | Configuration | Run 1 | Run 2 | Run 3 | Run 4 | Run 5 | Avg. |
37
+ |-----------------|--------------|--------------|--------------|--------------|--------------|--------------|
38
+ | bs8-e10-lr5e-05 | [0.8191][1] | [0.8086][2] | [0.8237][3] | [0.8318][4] | [0.8235][5] | 82.13 ± 0.76 |
39
+ | bs8-e10-lr3e-05 | [0.8056][6] | [0.8183][7] | [0.8241][8] | [0.8431][9] | [0.8155][10] | 82.13 ± 1.24 |
40
+ | bs4-e10-lr5e-05 | [0.8055][11] | [0.822][12] | [0.8243][13] | [0.8093][14] | [0.8144][15] | 81.51 ± 0.72 |
41
+ | bs4-e10-lr3e-05 | [0.8039][16] | [0.8122][17] | [0.8073][18] | [0.8246][19] | [0.8132][20] | 81.22 ± 0.7 |
42
+
43
+ [1]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-1
44
+ [2]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-2
45
+ [3]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-3
46
+ [4]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-4
47
+ [5]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-5
48
+ [6]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-1
49
+ [7]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-2
50
+ [8]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-3
51
+ [9]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-4
52
+ [10]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-5
53
+ [11]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-1
54
+ [12]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-2
55
+ [13]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-3
56
+ [14]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-4
57
+ [15]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-5
58
+ [16]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-1
59
+ [17]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-2
60
+ [18]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-3
61
+ [19]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-4
62
+ [20]: https://hf.co/hmbench/hmbench-icdar-nl-hmbert-bs4-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-5
63
+
64
+ The [training log](training.log) and TensorBoard logs are also uploaded to the model hub.
65
+
66
+ More information about fine-tuning can be found [here](https://github.com/stefan-it/hmBench).
67
+
68
+ # Acknowledgements
69
+
70
+ We thank [Luisa März](https://github.com/LuisaMaerz), [Katharina Schmid](https://github.com/schmika) and
71
+ [Erion Çano](https://github.com/erionc) for their fruitful discussions about Historic Language Models.
72
+
73
+ Research supported with Cloud TPUs from Google's [TPU Research Cloud](https://sites.research.google/trc/about/) (TRC).
74
+ Many Thanks for providing access to the TPUs ❤️