Search Results
6/30/2025, 3:41:58 AM
>>105749511
you want an LLM with "vision" capabilities. you can go the ollama route or you can use joy caption, both have comfyui nodes. I had decent success with minicpm-v (5.5gig) but it's very basic, doesn't know a thing about artists. it's ok tho for upscaling when you can't be bothered to write a prompt. joycaption is a whole lot better esp. for niche/nsfw content, 15ish gb tho.
you want an LLM with "vision" capabilities. you can go the ollama route or you can use joy caption, both have comfyui nodes. I had decent success with minicpm-v (5.5gig) but it's very basic, doesn't know a thing about artists. it's ok tho for upscaling when you can't be bothered to write a prompt. joycaption is a whole lot better esp. for niche/nsfw content, 15ish gb tho.
Page 1