utils/data_loader.py

from datasets import load_dataset
dataset = load_dataset("subashdvorak/tiktok-formatted-story-v2", revision="embedded")
data = dataset['train'].add_faiss_index('embeddings')

def load_influencer_data():
    return data