How to use fenguhao/ngram-tokenizer with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("fenguhao/ngram-tokenizer", dtype="auto")
