Can I replace it with a trained SentencePiece model?

[mumumu09chi]

example：spm_tokenizer = MLukeTokenizer.sp_model("./model/sentencepiece.bpe.model")
Is there a way to change it?

[ryokan0123]

We haven't tried that.
It seems to be in the scope of the transformers library so I recommend referring to other related documents.