Diffusion Models: How AI Creates Art from Noise
6d ago · 12 min read · TLDR: Diffusion models work by first learning to add noise to an image, then learning to undo that noise. At inference time you start from pure static and iteratively denoise into a meaningful image. They power DALL-E, Midjourney, and Stable Diffusio...
Join discussion
Breaking the One-Token Bottleneck: Why Diffusion Models for Code/Text May Redefine AI Generation (and My Take on the Future)
Nov 16, 2025 · 4 min read · Key References Inception Labs funding news (TechCrunch) https://techcrunch.com/2025/11/06/inception-raises-50-million-to-build-diffusion-models-for-code-and-text/ PAN world model: “A World Model for General, Interactable, and Long-Horizon World Si...
Join discussion
JoyCaption: The Open, Uncensored VLM That Will Supercharge Your Diffusion Models
Oct 27, 2025 · 3 min read · 📝 Quick Summary: JoyCaption is an open-source Visual Language Model (VLM) designed for image captioning. It aims to provide a free, uncensored alternative to existing models like ChatGPT, enabling the training and fine-tuning of diffusion models on ...
Join discussionDiffusion Model Overview
Oct 6, 2025 · 23 min read · Comprehensive Guide to Diffusion Models Table of Contents Guide to Diffusion Models Introduction to Diffusion Models How Diffusion Models Work Diffusion vs Autoregressive Models Comprehensive Model Comparison Trade-offs and Considerations Hybr...
Join discussionCreating an App that uses Diffusion Model to Generate Images
Jul 9, 2025 · 4 min read · Generative Modelling It is about teaching a model to generate new data that resembles its training data. There are numerous open access models that are trained on large datasets available for use. These pre-trained models can generate new data. Diffu...
Join discussion[Pixel Post] Riding the Noise: A 2025 Snapshot of State-of-the-Art Diffusion Models
Jul 7, 2025 · 10 min read · Diffusion models went from niche curiosity to the de facto workhorse of generative AI in less than five years. If you blinked during finals week, you probably missed three new sampling tricks and a Transformer architecture with better biceps than you...
Join discussion
[Pixel Post] Instant Pixel Animations: New Model Blueprint
Jul 6, 2025 · 3 min read · Goal: One-shot generation of coherent, stylized, multi-frame sprite animations (e.g. walk cycles, attacks) as either a horizontal grid or animated GIF.Target resolution: 64×64 per frame, 8-12 frames, full body character animations for 2D games. Step...
Join discussion
Flux.1 Control Net
Jun 27, 2025 · 2 min read · Testing Setup: Base Model: black-forest-labs/FLUX.1-dev LoRA Model: prithivMLmods/Retro-Pixel-Flux-LoRA (scale=0.8) ControlNet: Shakker-Labs/FLUX.1-dev-ControlNet-Union-Pro Prompt: 8 by 8 Retro Pixel Art Character Sheet of NPC characters for pixe...
Join discussion
KL Divergence
Jun 5, 2025 · 17 min read · Remember when we chatted about how AI models like Generative Adversarial Networks (GANs) learn to create? Well, there's another class of AI magicians : Variational Autoencoders (VAEs) and Diffusion Models, that achieve incredible feats of generation ...
Join discussion