from datasets import load_dataset, DatasetDict

# Define paths to data files
data_files = {
    "train": ["chunk_0.parquet", "chunk_1.parquet", "chunk_2.parquet"],
    "validation": ["chunk_3.parquet", "chunk_4.parquet", "chunk_5.parquet"],
}

# Load dataset with splits
dataset = load_dataset('parquet', data_files=data_files)