it is worth mentioning that the authors used a convolutional architecture to encode the condition image before feeding it within the cloned U-Net.
MULTIMODAL AI
-
Controlling Image Structure Generation: Latest Results
By
–
the way how this work enables us to control the structure from the images we generate is truly interesting. here are some of our favorite results so far:
-
ControlNet Architecture with Stable Diffusion Explained
By
–
the following is a representation of ControlNet's architecture when used with Stable Diffusion
-
ControlNet: Guide complet du fonctionnement et des applications
By
–
what is ControlNet and how does it work?
-
ControlNet: Conditioning Diffusion Models on Arbitrary Input Features
By
–
ControlNet is a method that can be used to condition diffusion models on arbitrary input features, such as image edges, segmentation maps, or human poses.
-
Kaggle Competition: Stable Diffusion Image-to-Prompts Challenge
By
–
Link to the competition https://
kaggle.com/competitions/s
table-diffusion-image-to-prompts/overview/prizes
… -
Stable Diffusion Kaggle Competition: Predict Image Prompts
By
–
Guess the prompt and win cash prizes Can you predict the text prompt from a generated image? If yes, then this power-packed AI competition from Stable Diffusion is for you. Participate now on Kaggle
-
Workshop on Generative Models Prompting at FITC Toronto
By
–
I'll be giving a workshop at FITC Toronto in April. Focus will be on interfacing with generative models of text and images, and making "prompting" more composable. Canada friends, hope you can make it!
-
Hugging Face Search Launches with 75,000 Searches in One Week
By
–
Wow, can you believe it's only been one week since https://t.co/OW9Eqe35zN launched and we've already had around 75,000 searches?! 🤯
— Vaibhav (VB) Srivastav (@reach_vb) 17 février 2023
It's amazing how effortless it is to find all sorts of incredible open-source models, datasets, and spaces across different modalities. 🤗 pic.twitter.com/Sg1m3oAWhWWow, can you believe it's only been one week since http://
hf.co/search launched and we've already had around 75,000 searches?! It's amazing how effortless it is to find all sorts of incredible open-source models, datasets, and spaces across different modalities. -
MultiDiffusion enables granular image generation control with segmentation maps
By
–
Similar a cómo funcionaba el trabajo de NVIDIA de hace unos meses, MultiDiffusion ofrece poder generar imágenes con un mayor control:
— Carlos Santana (@DotCSV) 17 février 2023
👉 Dibujando mapas de segmentación.
👉 y asociando cada mapa con un prompt!
Mola bastante ver cómo todas las herramientas se van sofisticando! https://t.co/TO1bROkd4WSimilar a cómo funcionaba el trabajo de NVIDIA de hace unos meses, MultiDiffusion ofrece poder generar imágenes con un mayor control: Dibujando mapas de segmentación. y asociando cada mapa con un prompt! Mola bastante ver cómo todas las herramientas se van sofisticando!