ResearchOct 19
OpenVLHarness: Real World Visual Agentic Tool-Use with Multimodal Memory
Upvotes97Authors11
Loading…
Glossary term
Able to process more than one type of input, such as text, images, audio and video.
A multimodal model handles more than one kind of data, such as text, images, audio and video, both as input and sometimes as output. It can describe a photo, answer questions about a chart or reply to speech.
Typically each type is turned into tokens or embeddings that one model processes together.
Most real-world information is not plain text, so multimodal models can help with far more tasks.
ResearchOct 19
Upvotes97Authors11
ResearchOct 14
Upvotes48Authors17