README.md

---
license: mit
language:
- fra
- ita
- por
- spa #['fra_Latn', 'ita_Latn', 'por_Latn', 'spa_Latn']  # ISO 639-3 code or "und" if not identifiable
tags:
- tokenizer
- bpe
- flexitok
- fineweb2
---

# Byte-Level BPE Tokenizer: ['fra_Latn', 'ita_Latn', 'por_Latn', 'spa_Latn'] (32K)

A **Byte-Level BPE** tokenizer trained on **['fra_Latn', 'ita_Latn', 'por_Latn', 'spa_Latn']** data from Fineweb-2-HQ.

## Training Details

| Parameter | Value |
|-----------|-------|
| Algorithm | Byte-Level BPE |
| Language | `['fra_Latn', 'ita_Latn', 'por_Latn', 'spa_Latn']` |
| Target Vocab Size | 32,000 |
| Final Vocab Size | 32,871 |
| Pre-tokenizer | custom:fra_Latn |
| Number handling | ltr_3digit |
| Contraction handling | True |
| Normalizer | NFC |
| Special Tokens | `<s>`, `</s>`, `<pad>`, `<unk>` |
| Training Shards | 8 |

## Usage

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_script_Roma_32000")
tokens = tokenizer.encode("Hello, world!")
```

## Files

- `tokenizer.json` — Full HuggingFace tokenizer
- `vocab.json` — Vocabulary mapping
- `merges.txt` — BPE merge rules

## Sample Encoding
| Text | Tokens | Token IDs |
|------|--------|-----------|
| `Hello, world! 12345 This is a test. こんにちは` | `H, ello, ,, Ġworld, !, Ġ, 123, 45, ĠThis, Ġis, Ġa, Ġtest, ., Ġ, ãģ, ĵ, ãĤ, ĵ, ãģ, «` | `42, 2110, 14, 25291, 3, 223, 22415, 4328, 17636, 1008, 267, 3037, 16, 223, 8090, 244, 14187, 244, 8090, 107` |