#!/usr/bin/env python3
from datasets import load_dataset
from tokenizers import ByteLevelBPETokenizer
# Load dataset
dataset = load_dataset('csv', data_files={'train': ["/home/nipunsadvilkar/mr_data/mr_train_df.csv"], "validation": ["/home/nipunsadvilkar/mr_data/mr_validation_df.csv"]}, split="train")
# Instantiate tokenizer
tokenizer = ByteLevelBPETokenizer()

def batch_iterator(batch_size=1000):
    for i in range(0, len(dataset), batch_size):
        yield dataset["text"][i: i + batch_size]

# Customized training
tokenizer.train_from_iterator(batch_iterator(), vocab_size=50265, min_frequency=2, special_tokens=[
    "<s>",
    "<pad>",
    "</s>",
    "<unk>",
    "<mask>",
])
# Save files to disk
tokenizer.save("./tokenizer.json")