Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
ControlNets for Virtual Staging
Exploring custom controlnet training for virtual staging: methods, architecture variations, and comparative results on converting empty rooms into furnished scenes.
I have been training several custom models for AI fully automated virtual staging of rooms. These allow the input of empty room images, and output a room with furniture. I would like to share several experiments (different architectures), and show the results of each.
Demonstrates virtual staging utilizing a parameter-efficient IC LoRA fine-tuning technique.
- ControlNetControlNet is a neural network architecture that adds precise spatial conditioning (e.g., Canny edges, OpenPose keypoints) to large, pretrained text-to-image diffusion models like Stable Diffusion.ControlNet is a game-changer for diffusion models, providing fine-grained control over image composition. It works by creating two copies of the model’s weights: a 'locked' copy to preserve the original Stable Diffusion capabilities and a 'trainable' copy for new conditions. This connection uses 'zero convolutions' to ensure no harmful noise affects the finetuning. This architecture allows training on small datasets (under 50k images) for tasks like pose-to-image or depth-to-image, efficiently steering the output with specific inputs like OpenPose keypoints or Canny edge maps.
Related projects
Generate precise and coherent images on your laptop with Stable Diffudion
Paris
This talk demonstrates controlling poses, styles, and character consistency using Stable Diffusion and ComfyUI for customized, inclusive image…
Agents for building bespoke machine learning models
Paris
Exploring how agents build custom machine learning models from customer data, covering toolkits, orchestration layer, template framework, and…
Infrahub - AI enabled road maintenance
New York City
Learn how AI vision models process street images to create and maintain a comprehensive road infrastructure database, enabling…
Space LLM
Paris
Learn how to generate architectural floor plans using fine‑tuned large language models, bridging text‑based AI with pixel‑based generative…
x2 faster diffusion model in 3 lines of code
Paris
This talk demonstrates how to use pruna to compress text-to-image diffusion models in three lines of code, doubling…
The workday AI health coach
Amsterdam
Learn how Mobi, an AI health coach, helps desk workers build healthy habits and prevent musculoskeletal disorders, reducing…