{
    "metadata": {
        "task_id": "TXT_INTER_00308",
        "task_level": "Level1",
        "task_type": "TXT_INTER",
        "original_video": "Data/Raw/00308/video.mp4",
        "original_motion": "Data/Raw/00308/human_motion.pkl",
        "duration": 14.36
    },
    "input": {
        "prompts": {
            "general_instruction": "You are an advanced 3D motion generation model designed to synthesize physically plausible human movements. I will provide you with semantic text inputs detailing the starting segment and the ending segment of a human action, with a temporal gap in between. These descriptions detail the kinematics of the person's whole body, hands, and legs for those specific segments. Your objective is to predict and generate the missing intermediate movements to bridge this temporal gap. You must ensure the generated motion logically connects the starting posture to the ending posture and maintains physical balance. Your output MUST be a continuous sequence of ACTION TOKENS representing the missing middle sequence.",
            "modifier_instruction": ""
        },
        "modalities": {
            "text": "Background: The person is a delivery worker in a yellow raincoat who accidentally drops two pizza boxes while walking, then pauses to retrieve and restack the items before continuing toward a door. Ends details: The person, wearing a yellow raincoat and a delivery backpack, is walking away from the camera on an asphalt surface, holding both hands in front of their torso to carry two flat white pizza boxes. The middle sequence is missing. The person then stands up from a squat, turns around, and continues walking away from the camera towards a door, with their hands holding the stacked white boxes firmly in front of their body.",
            "audios": [],
            "videos_processed": [],
            "spatial_coordinates": [],
            "input_images": []
        }
    },
    "ground_truth": {
        "motion_parameters": "Data/Processed/Text/TXT_INTER/00308_inter_gt.pkl",
        "rendered_video": "Data/Raw/00308/rendered.mp4"
    }
}