ireland-news-word-gap-predi.../regular_roberta_from_scratch/1_train_tokenizer.py
Jakub Pokrywka 51b778318d init
2021-11-04 11:37:51 +01:00

19 lines
383 B
Python

from pathlib import Path
from tokenizers import ByteLevelBPETokenizer
paths = ['../train/in_text_only.csv']
# Initialize a tokenizer
tokenizer = ByteLevelBPETokenizer()
# Customize training
tokenizer.train(files=paths, vocab_size=50265, min_frequency=2, special_tokens=[
"<s>",
"<pad>",
"</s>",
"<unk>",
"<mask>",
])
tokenizer.save_model("./tokenizer_model")