File size: 1,699 Bytes
35adce0
 
 
 
 
 
 
 
2b1d3c6
35adce0
 
 
2b1d3c6
35adce0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
---
language:
- de
- gmh
library_tags:
- transformers
- torch
tags:
- text-generation
- mt5
- middle-high-german
- normalization
pipeline_tag: text-generation
base_model: google/mt5-base
---

# normaere-model

Transformer-based seq2seq normalizer for Middle High German (MHG) text, fine-tuned from `google/mt5-base`.

## Model Details

- **Base Model:** google/mt5-base (580M parameters)
- **Architecture:** MT5ForConditionalGeneration (encoder-decoder)
- **Precision:** FP16
- **Max Sequence Length:** 512 tokens
- **Vocabulary Size:** 250,117 (augmented with medieval abbreviation characters)

## Intended Uses

Normalizes Middle High German texts (ca. 1050–1500) according to the standards of the Referenzkorpus Mittelhochdeutsch (ReM):

- Solves common abbreviations
- Splits off and expands common pro- and enclitics
- Applies editorial post-processing

## Training Data

Fine-tuned on the Referenzkorpus Mittelhochdeutsch (ReM), Version 2.1.

## Usage

```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("JonasHermann/normaere-model")
model = AutoModelForSeq2SeqLM.from_pretrained("JonasHermann/normaere-model")

input_text = "die stete wârheit"
inputs = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**inputs)
result = tokenizer.decode(outputs[0], skip_special_tokens=True)
```

## References

- mT5: [Xue et al., 2021](https://arxiv.org/abs/2010.11934)
- Training corpus: Roussel, Adam; Klein, Thomas; Dipper, Stefanie; Wegera, Klaus-Peter; Wich-Reif, Claudia (2024). Referenzkorpus Mittelhochdeutsch (1050–1350), Version 2.1, https://www.linguistics.ruhr-uni-bochum.de/rem/. ISLRN 937-948-254-174-0.