Upload folder using huggingface_hub

1fe3a86 verified 27 days ago

1.5 kB

license: mit
language:
  - dan
  - deu
  - nld
  - swe
tags:
  - tokenizer
  - bpe
  - flexitok
  - fineweb2

Byte-Level BPE Tokenizer: ['dan_Latn', 'deu_Latn', 'nld_Latn', 'swe_Latn'] (32K)

A Byte-Level BPE tokenizer trained on ['dan_Latn', 'deu_Latn', 'nld_Latn', 'swe_Latn'] data from Fineweb-2-HQ.

Training Details

Parameter	Value
Algorithm	Byte-Level BPE
Language	`['dan_Latn', 'deu_Latn', 'nld_Latn', 'swe_Latn']`
Target Vocab Size	32,000
Final Vocab Size	32,865
Pre-tokenizer	custom:dan_Latn
Number handling	ltr_3digit
Contraction handling	True
Normalizer	NFC
Special Tokens	`<s>`, `</s>`, `<pad>`, `<unk>`
Training Shards	8

Usage

from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_script_Germ_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«, ãģ`	`42, 13486, 14, 21745, 3, 223, 19219, 3832, 16775, 516, 270, 5190, 16, 223, 3768, 244, 5986, 244, 30698, 3768`