Pegasus 1.5 Model Now Available Through Amazon Bedrock
TwelveLabs' video language model, Pegasus 1.5, has launched on Amazon Bedrock, offering timestamped metadata generation for videos up to two hours long.
TwelveLabs and Amazon Web Services have announced the availability of the Pegasus 1.5 video language model, developed by TwelveLabs, on Amazon Bedrock. This model transforms long-form videos into structured, timestamped metadata output with the help of artificial intelligence.
Pegasus 1.5 can analyze entire videos to find segments matching user-defined criteria and return metadata in a structured JSON format. The release introduces innovations such as time-based metadata generation, video segmentation, multimodal prompting with reference images, on-the-fly processing, and expanded capacity for videos up to two hours long. Users can perform video analysis directly via an Amazon S3 URI or base64 string without the need for pre-indexing, keeping data within the AWS data boundary.
Sources
- aws.amazon.com · Link