diff --git a/data/processed/train_pretrain.parquet b/data/processed/train_pretrain.parquet index 31f0e0b..3bd3047 100644 Binary files a/data/processed/train_pretrain.parquet and b/data/processed/train_pretrain.parquet differ diff --git a/data/processed/val_pretrain.parquet b/data/processed/val_pretrain.parquet index 159c5ab..f9a4cc0 100644 Binary files a/data/processed/val_pretrain.parquet and b/data/processed/val_pretrain.parquet differ diff --git a/scripts/process_external.py b/scripts/process_external.py index b29229d..d2e4f28 100644 --- a/scripts/process_external.py +++ b/scripts/process_external.py @@ -18,7 +18,7 @@ app = typer.Typer() def main( input_path: Path = EXTERNAL_DATA_DIR / "all_data_LiON.csv", output_dir: Path = PROCESSED_DATA_DIR, - train_ratio: float = 0.9, + train_ratio: float = 0.7, seed: int = 42, ): """