llama 3.2

[Aevid]

is there a way to consistently get vector to take a picture and upload the image to the llm on together ai using the Llama 3.2 model? it does it very rarely on occasion but the more common outcome is the return of a fictional description of where it is

separately the model reads out the command to play animations rather than executing them when returning a response