bin/train_bpe.py

# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #
#                                                           #
#   This file was created by: Alberto Palomo Alonso         #
# Universidad de Alcalá - Escuela Politécnica Superior      #
#                                                           #
# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #
# Import statements:
from src import SegmentationTokenizer, SentenceSegmenter, SegmentationDataset
from bin.config import TokenizerConfig


# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #
#                        MAIN FUNC                          #
# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #
def main():
    # Get configs:
    wikipedia_dataset_path: str = input('Enter path to wikipedia dataset: ')
    config = TokenizerConfig()

    # Build tokenizer:
    tokenizer = SegmentationTokenizer(
        vocab_size=config.vocab_size,
        min_frequency=config.min_frequency,
        max_length=config.max_length
    )

    # Build segmenter:
    segmenter = SentenceSegmenter(max_sentences=config.max_sentences)

    # Build dataset:
    dataset = SegmentationDataset(
        wikipedia_dataset_path,
        tokenizer=tokenizer,
        segmenter=segmenter,
        percentage=config.dataset_portion
    )

    # Train tokenizer:
    tokenizer.train_from_iterator(
        tokenizer.build_iterator(dataset.huggingface_dataset, batch_size=config.train_batch_size)
    )
    tokenizer.save('tokenizer_32768.json')

    print('Tokenizer saved to tokenizer_32768.json')


# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #
#                        MAIN                               #
# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #
if __name__ == '__main__':
    main()


# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #
#                        END OF FILE                        #
# - x - x - x - x - x - x - x - x - x - x - x - x - x - x - #