---
license: mit
language:
- rus

tags:
- tokenizer
- bpe
- flexitok
- fineweb2
---

# Byte-Level BPE Tokenizer: ['rus_Cyrl'] (8K)

A **Byte-Level BPE** tokenizer trained on **['rus_Cyrl']** data from Fineweb-2-HQ.

## Training Details

| Parameter | Value |
|-----------|-------|
| Algorithm | Byte-Level BPE |
| Language | `['rus_Cyrl']` |
| Target Vocab Size | 8,000 |
| Final Vocab Size | 9,016 |
| Pre-tokenizer | custom:rus_Cyrl |
| Number handling | ltr_3digit |
| Contraction handling | False |
| Normalizer | NFC |
| Special Tokens | `<s>`, `</s>`, `<pad>`, `<unk>` |
| Training Shards | 2, ['fineweb_2_hq.rus_Cyrl.chunk.00.jsonl', 'fineweb_2_hq.rus_Cyrl.chunk.01.jsonl'] |

## Usage

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_ltr_rus_Cyrl_8000_v3")
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, ell, o, ,, Ġ, w, or, ld, !, Ġ, 123, 45, Ġ, Th, is, Ġ, is, Ġ, a, Ġ` | `42, 4444, 81, 14, 223, 89, 789, 5081, 3, 223, 8315, 4106, 223, 5200, 1143, 223, 1143, 223, 67, 223` |

Command used to create this tokenizer:
```bash
['/home/gsa/tokenizers2/flexitok/tokenizer_training/train_tokenizers.py', 'algorithm=bpe', 'vocab_size=8000', 'langs=[rus_Cyrl]', 'pretok_behavior=isolated', 'data_dir=/scratch/gsa/data/flexitok/', 'output_dir=/scratch/gsa/trained_tokenizers', 'pretokenizer=custom:rus_Cyrl', 'number_handling=ltr_3digit', 'add_numbers=true', 'handle_contractions=false', 'unicode_normalization=nfc', 'use_byte_level_regex=false', 'byte_fallback=false', 'strip_zero_width=false', 'cjk_char_split=false', 'cjk_char_coverage=0', 'add_cjk_chars=false', 'max_han_run=-1', 'max_lines=500_000', 'hf.publish_to_hf=true', 'hf_repo_prefix=flexitok/', 'hf.hf_repo_id=flexitok/bpe_ltr_rus_Cyrl_8000_v3', 'hf.collections=[flexitok/fineweb2-hq-tokenizers-v3,flexitok/8000-vocab-v3]']