---
language:
- sv
widget:
- text: "Den i HandelstidniDgens g&rdagsnnmmer omtalade hvalfisken, sorn fångats i Frölnndaviken"
  example_title: "News article #1"
- text: "En Gosse fur plats nu genast ! inetallyrkc, JU 83 Drottninggatan."
  example_title: "News article #2"
- text: "AfgäiigStiden bestämmes wid fartyget» hltkomst."
  example_title: "News article #3"
- text: "Elt godt Fortepiano om 6 octaver, ifräân Contra-F till o< med fyrſtrukna F, förfäljes af underte>uad för 260 R:dr Rgs."
  example_title: "Long-s piano ad"
---

# Swedish OCR correction

<!-- Provide a quick summary of what the model is/does. -->

This model corrects OCR errors in Swedish text.

## Try it!
- On short texts in the inference widget to the right -> 
- On files or longer texts in the [demo](https://huggingface.co/spaces/viklofg/swedish-ocr-correction-demo)

## Model Description

This model is a fine-tuned version of [byt5-small](https://huggingface.co/google/byt5-small), a character-level multilingual transformer.
The fine-tuning data consists of OCR samples from Swedish newspapers and historical documents.
The model works on texts up to 128 UTF-8 bytes (see [Length limit](#length-limit)).

<!-- ### Model Description-->

<!-- Provide a longer summary of what this model is. 

- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]

### Model Sources [optional]

- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]-->


## Training Data

The base model byt5 is pre-trained on [mc4](https://huggingface.co/datasets/mc4). This fine-tuned version is further trained on:

- Swedish newspapers from 1818 to 2018. Parts of the dataset are available from Språkbanken Text: [Swedish newspapers 1818-1870](https://spraakbanken.gu.se/en/resources/svenska-tidningar-1818-1870), [Swedish newspapers 1871-1906](https://spraakbanken.gu.se/resurser/svenska-tidningar-1871-1906).
- Swedish blackletter documents from 1626 to 1816, available from Språkbaknen Text: [Swedish fraktur 1626-1816](https://spraakbanken.gu.se/resurser/svensk-fraktur-1626-1816)

This data includes characters not used in Swedish today, such as the long s (ſ) and the esszett ligature (ß), which means that the model should be able to handle texts with these characters.
See for example the example titled _Long-s piano ad_ in the inference widget to the right.

## Usage
Use the code below to get started with the model.

```python
from transformers import pipeline, T5ForConditionalGeneration, AutoTokenizer

model = T5ForConditionalGeneration.from_pretrained('viklofg/swedish-ocr-correction')
tokenizer = AutoTokenizer.from_pretrained('google/byt5-small')
pipe = pipeline('text2text-generation', model=model, tokenizer=tokenizer)

ocr = 'Den i HandelstidniDgens g&rdagsnnmmer omtalade hvalfisken, sorn fångats i Frölnndaviken'
output = pipe(ocr)
print(output)
```

### Length limit
The model accepts input sequences of at most 128 UTF-8 bytes, longer sequences are truncated to this limit. 128 UTF-8 bytes corresponds to slightly less than 128 characters of Swedish text since most characters are encoded as one byte, but non-ASCII characters such as Å, Ä, and Ö are encoded as two (or more) bytes.