Yesterday @GoogleAI released Flan T5, a model that can solve 1800 different tasks 🤯
Since then
- Open-source models: https://t.co/X6q4mLouR0
- Support in @huggingface transformers
- An interactive demo I created to play with the model https://t.co/49e6MkN14E
This is not a drill. I was able to run #stablediffusion locally on an iPhone XS. Not via server. On. My. Phone. The prompt: “pineapple on a white table.”Watch it transform. (I broke the model into about 20 CoreML models to avoid running out of memory
What do @StabilityAI@EMostaque#stablediffusion & @BigscienceW Bloom - aka the coolest new models ;) - have in common?
They both use a new gen of ML licenses aimed at making ML more open & inclusive while keeping it harder to do harm with them. So cool!
https://t.co/o5ELwjc6vk
Top 20 dead artists used in @StableDiffusion Discord bot queries as of today, 1-2 PM, among the 773 artists in the collective archive of @sureailabs, @proximasan, @KyrickYoung and @EErratica. Can't wait for the Disco Diffusion 4000+ artists to be added !
Be sure to add time to your schedule to visit the #SIGGRAPH2022 Posters to experience an interactive forum and to hear presentations about the latest developments and ongoing #research in the industry. Posters is open today until 5:30 pm PDT in Vancouver. https://t.co/6Tf5P2Lba4
these are what I call 'latent paintings'...
basically a GAN interpolation based on a dataset made up of thousands of AI images based on the same prompt - with slight random variations.
Developing a new image and animation.
If it works, it'll be a part of our sound-visual performance 😎
- Generating basic images in @WOMBO (language-driven + based on influencing image)
- Processing them in @Photoshop@NeuralFilters
- To be cont... ))
#NFTCommumity#AIart
the difference some basic coloring adds is nutso tho. alien added with #dalle whole image composited and refined in photoshop. left is color-corrected, right is one or two stages more raw