Datasets:

Languages:
English
ArXiv:
License:
TimeIT / data /video_highlight_detection /video_highlight_detection_instructions.json
ShuhuaiRen's picture
upload data
7457806
raw
history blame
1.6 kB
{
"0": "You are given a video from the QVHighlights dataset. Please find the highlight contents in the video described by a sentence query, determining the highlight timestamps and its saliency score on a scale from 1 to 5. The output format should be like: 'The highlight timestamps are in the 82, 84, 86, 88, 90, 92, 94, 96, 98, 100 seconds. Their saliency scores are 1.3, 1.7, 1.7, 1.7, 1.7, 1.3, 1.7, 2.3, 2.3, 2.3'. Now I will give you the sentence query: <query_placeholder>. Please return the query-based highlight timestamps and salient scores.",
"1": "Watch the provided video and mark out the scenes that stand out based on the description: <query_placeholder>. Document the timestamps of these highlights and evaluate their saliency scores.",
"2": "Perform a thorough review of the video content, extracting key highlight moments that align with <query_placeholder>. It is essential to record the times of these moments and assign a distinct saliency value to each.",
"3": "Examine the video and, in accordance with query <query_placeholder>, highlight the standout moments. You're required to provide the exact timing alongside a saliency rating for each segment.",
"4": "In the video presented, seek moments that are a perfect match with <query_placeholder>. It's vital to notate their timestamps and to score each based on their level of saliency.",
"5": "Go through the video content, and upon identifying highlight moments that resonate with <query_placeholder>, list their timestamps. Subsequently, provide a saliency score for each identified highlight."
}