8d43da9a82fd73db14c4cacfa068261d84ea2610,scripts/bert/run_pretraining.py,,,#,383

Before Change

if __name__ == "__main__":
    random_seed = random.randint(0, 1000)
    nlp.utils.mkdir(args.ckpt_dir)

    dataset_name, vocab = args.dataset_name, None
    if args.sentencepiece:
        logging.info("loading vocab file from sentence piece model: %s", args.sentencepiece)

After Change

                                              args.max_predictions_per_seq)
        else:
            shuffle = True
            logging.info("args.num_buckets: {}, num_workers: {}, rank: {}".format(args.num_buckets,
                                                                                  num_workers,
                                                                                  rank))
            data_train = get_dataset_fn(args.data, batch_size,
                                        len(ctxs), shuffle, args.num_buckets, vocab,
                                        num_parts=num_workers, part_idx=rank,

Instances