is there a way to explore LAION/CLIP so that we can more quickly uncover the key words?
MULTIMODAL AI
-
Dreambooth Model Training Results with 22 Dog Photos
By
–
I trained dreambooth on just 22 pictures of my pup and I'm amazed by the results!
-
New Fine-Tuned Model Generates Emojis Automatically
By
–
A new fine-tuned model to generate Emojis is out!
-
CLIPSeg: Text-Based Image Inpainting with AI
By
–
Use text to inpaint objects in your images with CLIPSeg.https://t.co/PKtKM57SGN
— KREA AI (@krea_ai) 23 novembre 2022Use text to inpaint objects in your images with CLIPSeg.
-
Free Versatile Diffusion Demo Now Available on Hugging Face
By
–
Free Versatile Diffusion Demo available at @huggingface spaces.https://t.co/SFr2mSRdF4
— KREA AI (@krea_ai) 23 novembre 2022Free Versatile Diffusion Demo available at @huggingface spaces.
-
MagicVideo: Text-to-Video Generation Framework Using Diffusion
By
–
MagicVideo is an efficient text-to-video generation framework based on latent diffusion models that can produce photorealistic video clips from text descriptions.https://t.co/SUEk3csiB7 pic.twitter.com/MqVcOIwNY0
— KREA AI (@krea_ai) 23 novembre 2022MagicVideo is an efficient text-to-video generation framework based on latent diffusion models that can produce photorealistic video clips from text descriptions. https://
magicvideo.github.io -
Versatile Diffusion: Open-Source Multimodal AI Model Released
By
–
Versatile Diffusion can handle text-to-image, image-to-text, image-variation, and text-variation tasks. and the code is open-source! https://
github.com/SHI-Labs/Versa
tile-Diffusion
… -
NVIDIA SPACEx: Controllable High-Quality Portrait Animation Technology
By
–
Moody Moving Faces: NVIDIA’s SPACEx Delivers High-Quality Portrait Animation with Controllable Expression https://
syncedreview.com/2022/11/22/moo
dy-moving-faces-nvidias-spacex-delivers-high-quality-portrait-animation-with-controllable-expression/
… -
Data efficiency in multimodal learning for navigation tasks
By
–
I would say from the notion of data efficiency or minimal amount of cofounders behind your data. Like using text only to learn some navigation tasks (in big bench) sounds suboptimal to me But yeah of course science is fractal and everything can be considered scientific somehow:)
-
Internet Data Lottery: Web Availability Driving AI/ML Development
By
–
the « internet data lottery »: improvements in AI/ML coming from the path of largest web data availability rather than most scientifically sound direction modalities (text) win the lottery when they happen to be the widest available on the web even if sub-optimal for some tasks