| Parameter | Value |
|---|---|
| Algorithm | Byte-Level BPE |
| Language | ces_Latn |
| Target Vocab Size | 4,000 |
| Final Vocab Size | 5,060 |
| Pre-tokenizer | custom:ces_Latn |
| Number handling | ltr_3digit |
| Contraction handling | True |
| Normalizer | NFC |
| Special Tokens | <s>, </s>, <pad>, <unk> |
| Training Shards | 2 |
1from transformers import AutoTokenizer
2
3tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_ltr_ces_Latn_4000_v2")
4tokens = tokenizer.encode("Hello, world!")tokenizer.json — Full HuggingFace tokenizervocab.json — Vocabulary mappingmerges.txt — BPE merge rules| Text | Tokens | Token IDs |
|---|---|---|
Hello, world! 12345 This is a test. こんにちは | H, el, lo, ,, Ġw, or, ld, !, Ġ, 123, 45, ĠT, h, is, Ġis, Ġa, Ġtest, ., Ġ, ã | 42, 632, 309, 14, 979, 327, 2468, 3, 223, 5011, 4184, 390, 74, 527, 2780, 272, 2445, 16, 223, 162 |