import pandas as pd df = pd.read_parquet('d:/Research/Rl4Phyx/ZeroSearch/One-Shot-RLVR/data/train/physics_vlm/metaphyx_all.parquet') print(f'Rows: {len(df)}') print(f'Columns: {df.columns.tolist()}') print(f'First answer: {df.iloc[0]["answer"]}') print(f'\nSample prompt:') print(df.iloc[0]['prompt'][0]['content'][:300])