Projects sharing features with VIDM

Official code for "Conditional Generation from Unconditional Diffusion Models using Denoiser Representations" (BMVC 2023)

ICML 2023 official implementation for "Input Perturbation Reduces Exposure Bias in Diffusion Models"

Representation Learning with Diffusion Models

Official code for "DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Steps" (Neurips 2022 Oral)

CVPR 2024: Residual Denoising Diffusion Models

DiT is a latent diffusion model and transformer-based generative AI framework implemented in PyTorch. It functions as a class-conditional image generator that replaces traditional convolutional backbones with a transformer architecture to synthesize high-fidelity images. The project utilizes patch-based latent processing and latent space compression to operate on low-dimensional image representations. It incorporates class-conditional guidance and adjustable guidance scales to control the visual content of generated images during the sampling process. The framework covers distributed model t

Elucidating The Design Space of Classifier-Guided Diffusion Generation

Speed up Stable Diffusion with this one simple trick!

Code for the Paper "Improving Diffusion Model Efficiency Through Patching"

"Applying Regularized Schrödinger-Bridge-Based Stochastic Process in Generative Modeling"

Official implementation of Cold-Diffusion for different transformations in pytorch.

Official implementation for Estimating the Optimal Covariance with Imperfect Mean in Diffusion Probabilistic Models (ICML 2022), and a reimplementation of Analytic-DPM: an Analytic Estimate of the Optimal Reverse Variance in Diffusion Probabilistic Models (ICLR 2022)

ICML 2023 Reflected Diffusion Models (https://arxiv.org/abs/2304.04740)

Diffusion Probabilistic Modeling for Video Generation

PyTorch implementation for "Parallel Sampling of Diffusion Models", NeurIPS 2023 Spotlight

Diffusion Models with Deterministic Normalizing Flow Priors

ICLR'23 Spotlight gDDIM: analyze and accelerate general diffusion models, isotropic and non-isotropic