I suppose it's the pedant in me that wants to point out that LLMs are used to augment labeled image models, typically diffusion ones, to produce generated imagery, rather than creating imagery with the LLM itself 🙃
I suppose it's the pedant in me that wants to point out that LLMs are used to augment labeled image models, typically diffusion ones, to produce generated imagery, rather than creating imagery with the LLM itself 🙃