Groq model info (vision) is incorrect for some models

[FlorianSchneiderIU]

There are multiple models that are actually capable of doing vision, but have vision = False, e.g.

"llama-3.2-11b-vision-preview": {
"vision": False,
"function_calling": False,
"json_output": False,
"family": "meta-llama3.2",
},
"llama-3.2-90b-vision-preview": {
"vision": False,
"function_calling": False,
"json_output": False,
"family": "meta-llama3.2",
},

[vballoli]

Thank you for highlighting this @FlorianSchneiderIU , I appreciate your time on this. I'll get to it as soon as possible