Replies: 2 comments 1 reply
-
There is a project called Visual Chat GPT made by I believe a Microsoft team member that has a pretty impressive demo in the ReadMe. I haven't compiled it or tried it out, but the tool is in Github to for anyone to try it. The AI can generate images from text, respond to queries about an image,modify an image, basically it's like DallE and GPT4 combined, and hyper charged. Sorry i don't have the link handy but a simple search should bring it up, or it's one of the most recent starred repos on my profile. |
Beta Was this translation helpful? Give feedback.
-
When will the ability be released to the public? Is there an ETA? |
Beta Was this translation helpful? Give feedback.
-
We have a compositional VQA dataset at which all current state-of-the-art vision-language models are performing below chance levels. The documentation for OpenAI Evals doesn't seem to mention anything about querying GPT-4 with imagery—is this ability available yet?
Beta Was this translation helpful? Give feedback.
All reactions