Testing OCR Post-Correction with GPT Models

Raw data should be unpacked into the data/raw folder.
To run GPT, create a file named ./env and add the line MY_API_KEY= (place your API key after the equals sign).

• Preparation of training data based on raw data is performed in the notebook training_data.ipynb.
• Fine-tuning of the models was conducted on the OpenAI platform.
• Evaluation is performed in the notebook eval.ipynb.