cff-version: 1.2.0
title: "Khmer Graph-Regularized Tokenizer v2.2-RC"
message: "If you use this model, please cite it as below."
type: software
authors:
  - family-names: "khopilot"
    given-names: "Niko"
    orcid: "https://orcid.org/0000-0000-0000-0000"
repository-code: "https://github.com/khopilot/tokkonizer-km"
url: "https://huggingface.co/khopilot/khmer-tokenizer-v7"
version: "2.2-rc"
date-released: "2025-10-07"
license: MIT
keywords:
  - khmer
  - tokenization
  - graph-regularization
  - nlp
  - semantic-embeddings
abstract: "A SentencePiece tokenizer with graph-regularized lexeme embeddings for Khmer. Achieves 43.25% Coherence@10 (+110% vs baseline) by structuring the semantic space according to morphological and synonymy relationships."