@julian_ai_art @fabianstelzer@SALT_VERSE We’ve got many folks in our Discord talking about image and story generation if you’d like to join us! Also working on a free, open-source tool for building with these models.
Generating children’s books and graphic novels are early use case targets:
https://t.co/s7vLeooZsy
@MaxencePastor@f_j_j_ We’ve been using a technique (discovered accidentally) with Copilot we call “reverse troubleshooting.”
Add a section to the Troubleshooting page of your docs explaining a problem you’ve encountered, and let Copilot guess the solution.
@Money17251696@Scobleizer Indeed! It's a technique that allows you to chain together the outputs of one prompt and use that as the input to another prompt.
One project that uses this for computer vision models is https://t.co/KNakvv95Rg
We're taking a similar approach, but for any type of model.
@f_j_j_ Even cooler! We hadn’t seen any worthwhile prompt injection defenses. Will definitely check that out. 🙌
We are still developing our first desktop release–we post design & dev updates to our Discord. Our plan is to integrate with every provider that we can.
Dust is Live🔥
As a first demo: a Dust app showcasing the now famous capability of models to generate code to answer math questions.
https://t.co/eANW67s5iB
(some comments in 🧵)
@goodside@gideonmann We’re working on a free, open-source tool to experiment with “prompt chaining” techniques. Intermediate revision of generations (and their subsequent prompts) are on the near-term roadmap.
Code generation and execution are also definitely on the agenda!
@hypnagogica@_akhaliq@huggingface@MokadyRon @hertz_amir It’s tough to keep up with! Luckily, I see @MokadyRon already answered your question.
Our community is building a resource to help keep track of this space as it develops: https://t.co/ZovAqk6zjZ
@yoheinakajima Very cool! Excited to check it out. 💪
We’re building a free, open-source desktop app that’s very similar. Also includes notions of “prompt chaining” & templates.
https://t.co/ZovAqk61ur
@jwegener And definitely check out Lexica if you haven’t already. The emerging technique of tokenizing specific snippets of text, which encode an entire visual style, is already pretty impressive.
https://t.co/V1zSnfQAjZ
Introducing Lexica – a search engine for AI-generated images and prompts.
Every image has a prompt and seed, so you can copy and remix anything for yourself.
Hopefully this makes Stable Diffusion prompting a bit less of a dark art and more of a science!
https://t.co/0YdmzHqqY0