data_juicer.ops.mapper.video_human_tracks_face_demographic_mapper module

class data_juicer.ops.mapper.video_human_tracks_face_demographic_mapper.VideoHumantrackFaceDemographicMapper(*args, **kwargs)

Bases: Mapper

Mapper to generate samples whose captions are generated based on a video-to-text model and sampled video frame.

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

__init__(original_data_save_path='./outputs/demographic', detect_interval: int = 5, tag_field_name: str = 'video_facetrack_attribute_demographic', *args, **kwargs)

Initialization method.

Parameters:

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

process_single(sample, rank=None)

For sample level, sample –> sample

Parameters:

sample – sample to process

Returns:

processed sample

find_median(int_list)

Finds the median of a list of integers.

Parameters:

int_list – List of integers.

Returns:

The median of the list.

most_frequent_element_ratio(str_list)

Finds the most frequent element in a list of strings and its proportion.

Parameters:

str_list – List of strings.

Returns:

A tuple (most_frequent_element, proportion).

top_3_emotions(emotion_list)

Calculates the average score for each emotion and returns the top 3 emotions with the highest average scores.

Parameters:

emotion_list – List of dictionaries, each containing seven emotions and their corresponding scores.

Returns:

List of tuples containing the top 3 emotions and their average scores.