1. Learn
    2. /
  2. Courses
    3. /
  3. Multi-Modal Models with Hugging Face
Connected

Exercise

Video audio splitting

In this exercise, you will be splitting a video advertising Bounce fabric softener from the 1980s. This is necessary so that we can isolate the image and audio inputs necessary for the individual image and audio classifier models.

Frames from the Bounce TV commercial

The ffmpeg_extract_subclip and VideoFileClip modules have been loaded. The MP4 file has been downloaded as bounce_ad.mp4.

Instructions

100 XP
  • Create a subclip file from bounce_ad.mp4 that begins at 0s and ends at 5s, with the output filename bounce_ad_5s.mp4.
  • Load the new subclip.
  • Extract the audio stream.
  • Write the audio stream to the filename bounce_ad_5s.mp3