Is training of LeoAI on images close to becoming cost effective or it is still ridiculous to go multimodal?
You are viewing a single comment's thread from:
Is training of LeoAI on images close to becoming cost effective or it is still ridiculous to go multimodal?
Slowly getting there. I have been toying with using AI Agents who can leverage ChatGPT (and similar models) to read images, then relay the info to LeoAI via threads
An interesting twist although there is an expense on ChatGPT for prompts. I dont know if the inference prices are dropping.