metadata
dataset_info:
features:
- name: video_id
dtype: string
- name: avg_aesthetic_score_laion_aesthetics
dtype: float64
- name: frame_aesthetic_scores_laion_aesthetics
dtype: string
- name: video_name
dtype: string
- name: motion_fb_motion_score
dtype: float64
- name: motion_lk_motion_score
dtype: float64
- name: frame_shot_categorization_shot_categorizer
dtype: string
- name: avg_subjects_subject_presence
dtype: float64
- name: avg_humans_subject_presence
dtype: float64
- name: avg_animals_subject_presence
dtype: float64
- name: frame_counts_subject_presence
dtype: string
- name: avg_vision_reward
dtype: float64
- name: frame_wise_rewards
dtype: string
- name: video_url
dtype: string
- name: scene_name
dtype: string
- name: SCENE_NUM
dtype: float64
- name: START_FRAME
dtype: float64
- name: END_FRAME
dtype: float64
- name: START_TIMECODE
dtype: string
- name: END_TIMECODE
dtype: string
- name: START_SECONDS
dtype: float64
- name: END_SECONDS
dtype: float64
- name: DURATION_SECONDS
dtype: float64
- name: GAP_FROM_PREVIOUS
dtype: float64
- name: prompt_x
dtype: string
- name: caption_t2v_style
dtype: string
- name: prompt_y
dtype: string
- name: caption_t2v_style_short
dtype: string
splits:
- name: train
num_bytes: 547236312
num_examples: 59933
download_size: 81242850
dataset_size: 547236312
configs:
- config_name: default
data_files:
- split: train
path: data/train-*