Slides for my talk at GitHub Universe 2022. See universe-22.vercel.app
↕ Use your arrow keys to navigate the slides.
Stuff mentioned in the slides:
🙏🏼 Request access to GPU Codespaces - GPU access in Codespaces is not yet generally available (as of November 9, 2022) so you'll need to request access using this form.
👩🚀 Stable Diffusion: The beloved open-source text-to-image model, running on Replicate. You can run it on Replicate using the web UI, or the API.
🀄️ Material Stable Diffusion: A fork of SD that generates tiling images
🎨 [Inpainter]https://github.com/replicate/inpainter): An open-source Next.js app that uses Replicate's API and Stable Diffusion to inpaint images. Check out the source code or try it out at https://inpainter.vercel.app.
🍿 Stable Diffusion animation: A model on Replicate that can animate Stable Diffusion by interpolating between two prompts
🧺 Replicate Collections - Curated sets of Replicate models that serve different purposes: Text to image, Image to text, Style transfer, Image restoration, etc.
📡 Stable Diffusion Textual Inversion: A demo repo illustrating how to run GitHub Codespaces to fine-tune Stable Diffusion with your own concept using textual inversion, and deploy it to Replicate as a cloud API.
🌏 Action Transcription from @simonw uses OpenAI Whisper and GitHub Actions to transcribe and translate YouTube videos. Just open a GitHub issue to fire it up!
📗 Once Upon a bot from @cbh123 uses GPT-3 and Stable Diffusion to create an original illustrated story.
⛓ Replicate CLI from @jamiesteven is an open-source command-line interface for running models on Replicate, and piping the output of one model as input into another.
These people are creating some really interesting models on Replicate:
To run this app on your own machine, you'll need to download and install Node.js, then:
npm install
npm run dev