Upload folder using huggingface_hub

fed3066 verified 22 days ago

1.38 kB

license: mit
language:
  - ind
  - vie
tags:
  - tokenizer
  - bpe
  - flexitok
  - fineweb2

Byte-Level BPE Tokenizer: ['ind_Latn', 'vie_Latn'] (16K)

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

Training Details

Parameter	Value
Algorithm	Byte-Level BPE
Language	`['ind_Latn', 'vie_Latn']`
Target Vocab Size	16,000
Final Vocab Size	16,959
Pre-tokenizer	custom:ind_Latn
Number handling	ltr_3digit
Contraction handling	True
Normalizer	NFC
Special Tokens	`<s>`, `</s>`, `<pad>`, `<unk>`
Training Shards	4

Usage

from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_script_SEAS_16000")
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, el, lo, ,, Ġw, orld, !, Ġ, 123, 45, ĠThis, Ġis, Ġa, Ġtest, ., Ġ, ãģ, ĵ, ã, Ĥ`	`42, 324, 2155, 14, 505, 4659, 3, 223, 16876, 4702, 15780, 1555, 1333, 8184, 16, 223, 11148, 244, 162, 227`