Analyze videos
The platform analyzes videos to generate text based on their content using a multimodal approach. This method analyzes the visuals, sounds, spoken words, and relationships between them. As a result, it provides a comprehensive understanding of your videos, capturing nuances that might be overlooked when using an unimodal interpretation.
Notes
- This feature is available only for the indexes that have the Pegasus video understanding engine enabled. See the Create an index section for details.
- Your prompts can be instructive or descriptive, or you can also phrase them as questions. For guidance on creating effective prompts, see the Prompt engineering page.
- The maximum length of a prompt is 2,000 tokens.
Procedure
Follow the steps in this guide to generate text based on the content of your videos:
-
From the Indexes page, find and select the index containing the videos based on which you wish to generate text.
-
Choose the Analyze tab.
-
Enter your prompt:
Note
You can also use one of the prompts that the Playground suggests.
-
(Optional) Temperature is a configurable parameter that controls the randomness of the text output generated by the model. A higher value generates more creative text, while a lower value results in more deterministic text output. Select the thermomether icon and then use the slider to tailor the behavior of the model to your requirements.
-
To analyze your video and generate text based on its content, select the icon at the bottom-right corner of the input box:
The Playground displays the generated text in the output panel:
-
(Optional) You can use the buttons displayed below the generated text to:
- Copy the generated text for use in other applications.
- View the code snippet the platform used to perform this request. You can copy and paste it into your application.
- Generate new text using the same parameters.
-
(Optional) To generate text for another video, use the Select another video button: