---
library_name: transformers
tags:
- legal
license: apache-2.0
datasets:
- byczong/pl-insurance-terms-struct
language:
- pl
base_model:
- naver-clova-ix/donut-base
pipeline_tag: image-text-to-text
---

# Model Card

Donut fine-tuned for full document structuring (parsing) on [pl-insurance-terms-struct](https://huggingface.co/datasets/byczong/pl-insurance-terms-struct) dataset.

Trained for 10 epochs with `max_seq_len=7168`.

- Field-level f1 score: 0.57
- TED-based accuracy: 0.67


Note: This model and its tokenizer were not (pre-) trained for Polish.