Update README
Browse files
README.md
CHANGED
@@ -2,6 +2,23 @@
|
|
2 |
datasets:
|
3 |
- aehrm/dtaec-lexica
|
4 |
language: de
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
---
|
6 |
|
7 |
# DTAEC Type Normalizer
|
@@ -31,7 +48,20 @@ model = AutoModelForSeq2SeqLM.from_pretrained('aehrm/dtaec-type-normalizer')
|
|
31 |
model_in = tokenizer(['Freyheit', 'seyn', 'selbstthätig'], return_tensors='pt', padding=True)
|
32 |
model_out = model.generate(**model_in)
|
33 |
|
34 |
-
print(tokenizer.batch_decode(model_out))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
```
|
36 |
|
37 |
|
|
|
2 |
datasets:
|
3 |
- aehrm/dtaec-lexica
|
4 |
language: de
|
5 |
+
pipeline_tag: translation
|
6 |
+
model-index:
|
7 |
+
- name: aehrm/dtaec-type-normalizer
|
8 |
+
results:
|
9 |
+
- task:
|
10 |
+
name: Historic Text Normalization (type-level)
|
11 |
+
type: translation
|
12 |
+
dataset:
|
13 |
+
name: DTA-EC Lexicon
|
14 |
+
type: aehrm/dtaec-lexica
|
15 |
+
metrics:
|
16 |
+
- name: Word Accuracy
|
17 |
+
type: accuracy
|
18 |
+
value: 0.9546
|
19 |
+
- name: Word Accuracy OOV
|
20 |
+
type: accuracy
|
21 |
+
value: 0.9096
|
22 |
---
|
23 |
|
24 |
# DTAEC Type Normalizer
|
|
|
48 |
model_in = tokenizer(['Freyheit', 'seyn', 'selbstthätig'], return_tensors='pt', padding=True)
|
49 |
model_out = model.generate(**model_in)
|
50 |
|
51 |
+
print(tokenizer.batch_decode(model_out, skip_special_tokens=True))
|
52 |
+
# >>> ['Freiheit', 'sein', 'selbsttätig']
|
53 |
+
```
|
54 |
+
|
55 |
+
Or, more compact using the huggingface `pipeline`:
|
56 |
+
|
57 |
+
```python
|
58 |
+
from transformers import pipeline
|
59 |
+
|
60 |
+
pipe = pipeline(model="aehrm/dtaec-type-normalizer")
|
61 |
+
out = pipe(['Freyheit', 'seyn', 'selbstthätig'])
|
62 |
+
|
63 |
+
print(out)
|
64 |
+
# >>> [{'generated_text': 'Freiheit'}, {'generated_text': 'sein'}, {'generated_text': 'selbsttätig'}]
|
65 |
```
|
66 |
|
67 |
|