toklens/fineweb_edu_12gb_tokenizer_sentencepiece

Byte-Level BPE Tokenizer: ['fw_edu'] (32K)

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

Training Details

Parameter	Value
Algorithm	Byte-Level BPE
Language	['fw_edu']
Target Vocab Size	32,000
Final Vocab Size	32,000
Pre-tokenizer	sentencepiece
Number handling	learned
Contraction handling	False
Normalizer	NFC
Special Tokens	<s>, </s>, <pad>, <unk>
Training Shards	2, ['fineweb_edu_10bt.chunk.00.jsonl', 'fineweb_edu_10bt.chunk.01.jsonl']

Usage

from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("None")
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	▁Hell, o,, ▁world, !, ▁12, 3, 45, ▁This, ▁is, ▁a, ▁test., ▁, こ, ん, に, ち, は	29987, 16984, 16629, 5, 17468, 23, 21086, 16370, 16032, 15979, 31029, 4548, 5337, 5401, 5361, 5351, 5365

Command used to create this tokenizer:

['/home/gsa/tokenizers2/flexitok/tokenizer_training/train_tokenizers.py', 'algorithm=bpe', 'vocab_size=32_000', 'langs=[fw_edu]', 'data_dir=/scratch/gsa/data/toklens/tokenizer-training/', 'output_dir=/scratch/gsa/trained_tokenizers/toklens-sentencepiece', 'pretokenizer=sentencepiece', 'number_handling=learned', 'handle_contractions=false', 'unicode_normalization=nfc', 'use_byte_level_regex=false', 'strip_zero_width=false', 'max_lines=-1', 'hf.publish_to_hf=true']