Multimodal architecture for building visual generative AI.
NVIDIA Edify is a multimodal architecture for developing visual generative AI models for image, 3D, 360 HDRi, physically based rendering (PBR) materials, and video. Using NVIDIA AI Foundry, service providers can train and customize Edify models to build commercially viable visual services built on top of NVIDIA NIM™.
Image courtesy of Getty Images
Create custom visual generative AI models for your applications and services. Choose NVIDIA's multimodal Edify architecture or seamlessly deploy the model of your preference.
Train the state-of-the-art NVIDIA Edify diffusion model on your data to generate a range of visual content types from 2D to 3D.
Build models to generate, control, and modify 4K images.
Generate 4K images with text or a reference image with greater prompt adherence and generation controls.
Add characters or objects with text prompts or replace specific elements.
Expand images to various aspect ratios with text prompts.
Reimagine a composition with a variety of characters and environments.
Transform sketches into unique images with guided prompts.
Segment any object automatically to generate masks quickly during editing workflows.
Transform a 1K image into a 4K image.
Develop models to accelerate the creation of 3D assets. Edify 3D includes a clean quads-based topology ready for editing and support for 4K PBR materials with de-lit base color, metal, and roughness maps.
Create simple 3D assets for rapid prototyping and set dressing.
Quickly generate 3D assets from a single image reference.
Train a model to create lighting for 3D scenes with 360-degree domes as environment maps. Create immersive environments with optimal PBR lighting and simple backplates.
Create environment maps with text or image prompts to relight 3D scenes.
Train models for PBR material generation.
Generate tileable materials and textures with text prompts—like brick or mosaic—for virtual scenes.
Train the Edify model to generate video.
Generate high-fidelity videos with temporal consistency.
Optimized solution for customizing commercially safe, visual generative AI models.
Pioneers have created commercially safe AI generators for images and 3D, revolutionizing creative workflows across industries with ongoing feature enhancements.
Advanced AI generators combine Getty Images’ pre-shot library for commercially safe and legally protected images with NVIDIA’s cutting-edge AI, enabling the rapid creation of high-quality imagery from text or image prompts.
Advanced editing tools, such as inpainting and outpainting, allow for rapid image modification. Customization capabilities for model fine-tuning, depth, segmentation, and sketch empower users to create visual content tailored to their preferences.
The commercially safe service also compensates artists and creators for allowing their work to be part of the model.
Shutterstock’s 3D AI service lets both experienced and novice 3D creators quickly conceptualize and create assets for prototypes, set dressing, immersive environments, and more.
Trained on Shutterstock’s high-quality licensed data, it filters inappropriate content and compensates contributing artists for each asset generated. You can integrate it into creative applications or plug-ins using Shutterstock APIs.
Explore who’s deploying custom visual foundation models.