Introducing "Spectrally-Guided Diffusion Noise Schedules", accepted to #ICML2026! With @kiamada.
We propose noise schedules that follow the power spectrum of each image, improving pixel diffusion quality and requiring fewer denoising steps.
At test time, we conditionally sample
Rethinking image representations for autoregressive generation! Our new Spectral Image Tokenizer works in the image spectrum, where the coarse-to-fine representation has a more natural sequential interpretation
Our new paper, "Spectral Image Tokenizer", is on arXiv! We train a tokenizer on DWT coefficients
that enables autoregressive coarse-to-fine image generation, w/ applications to multiscale text-to-image, and text-guided editing.
w/ @kiamada, @msuhail153arxiv.org/abs/2412.09607
Generate high quality textures with single mesh LDMs!
#CVPR2024
Our *intrinsic* 3D diffusion models, trained on a single mesh, can generate texture variations, perform inpainting, and even transfer textures to different shapes.
single-mesh-diffusion.github.io
w/@twmitchel & @_machc
