here to participate and dominate in this tournament :)
Capsum watches any video clip and generates captions in four distinct styles—formal, sarcastic, humorous-tech, and humorous-non-tech—by extracting scenes, sampling keyframes, and grounding each style in the same visual facts for accuracy and tone.
.png&w=256&q=75)
