bbpe/maddition_digit_lambda_voc_0_lambda_tok_0_lambda_seg_1_2000_minimal

Regularized BPE Tokenizer: ['digit'] (2K)

A Regularized BPE tokenizer trained on ['digit'] data from Fineweb-2-HQ.

Training Details

Parameter	Value
Algorithm	Regularized BPE
Language	['digit']
Target Vocab Size	2,000
Final Vocab Size	259
Pre-tokenizer	whitespace
Number handling	ltr_3digit
Contraction handling	False
Normalizer	NFC
Special Tokens	<s>, </s>, <pad>, <unk>
Training Shards	2, ['train.chunk.00.jsonl', 'val.chunk.00.jsonl']

Usage

from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("bbpe/maddition_digit_lambda_voc_0_lambda_tok_0_lambda_seg_1_2000_minimal")
tokens = tokenizer.encode("Hello, world!")

Files

• tokenizer.json — Full HuggingFace tokenizer
• vocab.json — Vocabulary mapping
• tokenizer.model — SentencePiece protobuf (if available)

Sample Encoding

Text	Tokens	Token IDs
22+9=31\nyirmi iki+dokuz=otuz bir\ntwenty two+nine=thirty one	2, 2, +, 9, =, 3, 1, \, n, y, i, r, m, i, i, k, i, +, d, o	20, 20, 13, 27, 31, 21, 19, 62, 80, 91, 75, 84, 79, 75, 75, 77, 75, 13, 70, 81

Command used to create this tokenizer:

['/home/gsa/tokenizers2/flexitok/tokenizer_training/train_tokenizers.py', 'algorithm=rbpe', 'vocab_size=2000', 'langs=[digit]', 'data_dir=/scratch/gsa/data/multilingual-addition/', 'output_dir=/scratch/gsa/trained_tokenizers/multilingual_addition', 'pretokenizer=whitespace', 'number_handling=ltr_3digit', 'add_numbers=false', 'handle_contractions=false', 'unicode_normalization=nfc', 'use_byte_level_regex=false', 'byte_fallback=false', 'strip_zero_width=false', 'cjk_char_split=false', 'lambda_voc=0', 'lambda_tok=0', 'lambda_seg=1', 'add_cjk_chars=false', 'max_lines=-1', 'test_string=22+9=31\\nyirmi iki+dokuz=otuz bir\\ntwenty two+nine=thirty one', 'hf.publish_to_hf=true', 'hf_repo_prefix=bbpe/', 'hf.hf_repo_id=bbpe/maddition_digit_lambda_voc_0_lambda_tok_0_lambda_seg_1_2000_minimal', 'hf.collections=[bbpe/multilingual_addition_tokenizers]']