search

Q-Future / Q-Bench

①[ICLR2024 Spotlight] (GPT-4V/Gemini-Pro/Qwen-VL-Plus+16 OS MLLMs) A benchmark for multi-modality LLMs (MLLMs) on low-level vision and visual quality assessment.
https://q-future.github.io/Q-Bench/
Other
242 stars 12 forks source link

Questions about gpt-4v-vs-human.png #7

Closed HenryHZY closed 10 months ago

HenryHZY commented 10 months ago

Hi @teowu, thanks for your great project.

I would like to ask a side question. Are your images https://github.com/Q-Future/Q-Bench/blob/master/gpt-4v-vs-human.png generated by an AI model? That is very vivid.

I'm not familiar with the diffusion model, so I would like to know what model and prompt you use. Thanks in advance:)

teowu commented 10 months ago

That is generated by DALLE-3. Prompts like 'GPT-4V doing an exam called "Q-Bench", in competition with a girl in glasses' may work.

HenryHZY commented 10 months ago

Thanks for your quick reply!