data_juicer.ops.mapper.video_captioning_face_attribute_emotion_mapper module

class data_juicer.ops.mapper.video_captioning_face_attribute_emotion_mapper.VideoCaptioningFaceAttributeEmotionMapper(*args, **kwargs)

基类：Mapper

Generate facial attribute and emotion descriptions for each person tracked in a video using a video-to-text model.

Source: This operator is a part of HumanVBench (CVPR 2026).

__init__(face_track_query: str = "Please describe the person's facial expression, tell me the person's emotion through the video, like Happiness, Excitement, Love, Gratitude, Relief, Pride, Anger, Sadness, Fear, Guilt, Shame, Disgust, Surprise, Confusion, Curiosity, Boredom ...", trust_remote_code: bool = False, cropping_face_video_temp_path='./temp_video_path', video_describe_model_path: str = 'DAMO-NLP-SG/VideoLLaMA3-7B', video_facetrack_attribute_emotion: str = 'video_facetrack_attribute_emotion', *args, **kwargs)

Initialization method.

参数:

hf_video_blip -- video-blip model name on huggingface to generate caption

Source: This operator is a part of HumanVBench (CVPR 2026).

process_single(sample, rank=None)

For sample level, sample --> sample

参数:

sample -- sample to process

返回:

processed sample