Image only input
#15
by
fedric95
- opened
Great work. From the doc it is not clear to me if llama guard 4 it is also meant to be used as image-only safety classifier (prompt with only images) or not (like llama guard vision)
Hi there, llama guard 4 is not meant to be a image-only safety classifier.
Great
No description provided.
Analyse
Hi there, llama guard 4 is not meant to be a image-only safety classifier.
Could you please clarify what is meant by "image-only" in this context?
Specifically, if the input includes a single image and a system prompt (but no user query), would this still be considered "image-only" by the model, or is it sufficient for performing multimodal safety classification?