Motion Vision
Streamlit
Video Captioning
In this project, the Video url is given to the AI model. The AI model is connected to LLM using FireWorks AI. There are many models which natively supports vision. This is our first submission in which we use a simple method in which we just fed the video url to the AI. and It reads and generate captions according to the System prompts and because the system prompt is our main Product so we have encrypted It so nobody can copy it. This a good security measure. We first tried to pre process the video so it just don't completely fill the context window. But later found that llm have their own preprocessor which converts the video into 360p and 1 fps. So this is our simple attempt
Category tags:Explore more applications
.png&w=96&q=75){kind=small}
Upcoming AI Hackathons
For Innovators & Creators
For Innovators & Creators
Explore more applications