[FEATURE] support ollama multimodal

[Eyalm321]

Describe the feature you'd like Right now, to have the llm recognize an image we have to use expansive apis, please support base64 image, Its already support behind the scenes by langchain. basically all we have to do to get this functionality is llm.generate(baseUrl, { images: ["base64_string"]}) and it will work with llava model, I tried node script with langchain seems easy enough to implement Ollama Llava Tool or Adjust the legacy ollama for it

Additional context Add any other context or screenshots about the feature request here.