---
license: mit
language:
- fas  # ISO 639-3 code or "und" if not identifiable
tags:
- tokenizer
- bpe
- flexitok
- fineweb2
---

# Byte-Level BPE Tokenizer: fas_Arab (32K)

A **Byte-Level BPE** tokenizer trained on **fas_Arab** data from Fineweb-2-HQ.

## Training Details

| Parameter | Value |
|-----------|-------|
| Algorithm | Byte-Level BPE |
| Language | `fas_Arab` |
| Target Vocab Size | 32,000 |
| Final Vocab Size | 32,808 |
| Pre-tokenizer | custom:fas_Arab |
| Number handling | ltr_3digit |
| Contraction handling | True |
| Normalizer | NONE |
| Special Tokens | `<s>`, `</s>`, `<pad>`, `<unk>` |
| Training Shards | 2 |

## Usage

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_ltr_fas_Arab_32000_v2")
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, ,, Ġ, world, !, Ġ, 123, 45, Ġ, This, Ġ, is, Ġ, a, Ġ, test, ., Ġ, ãģ` | `42, 25612, 14, 223, 22224, 3, 223, 18213, 5459, 223, 15727, 223, 1809, 223, 67, 223, 21816, 16, 223, 18685` |