---
license: cc-by-4.0
language:
- pl
library_name: transformers
tags:
- tokenizer
- fast-tokenizer
- polish
datasets:
- radlab/legal-mc4-pl
- radlab/wikipedia-pl
- radlab/kgr10
- clarin-knext/msmarco-pl
- clarin-knext/fiqa-pl
- clarin-knext/scifact-pl
- clarin-knext/nfcorpus-pl
---

This is polish fast tokenizer.

Number of documents used to train tokenizer:
 - 25 088 398


Sample usge with transformers:

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained('radlab/polish-fast-tokenizer')
tokenizer.decode(tokenizer("Ala ma kota i psa").input_ids)

```