ImageNet 64x-64

'ImageNet 64x64' is a benchmark dataset containing images from the ImageNet collection resized to 64x64 pixels, used to evaluate the performance of generative modeling approaches in image generation tasks.

Papers using ImageNet 64x-64 (13)

Simplified and Generalized Masked Diffusion for Discrete Data2024 · 3 citesSimplifying, Stabilizing and Scaling Continuous-Time Consistency Models2024 · 1 citesAlign Your Flow: Scaling Continuous-Time Flow Map Distillation2025Score-of-Mixture Training: Training One-Step Generative Models Made Simple via Score Estimation of Mixture Distributions2025VCT: Training Consistency Models with Variational Noise Coupling2025Consistency Models2023 · 26 citesLossy Compression with Gaussian Diffusion2022 · 20 citesOn Distillation of Guided Diffusion Models2022 · 8 citesWhy Are Conditional Generative Models Better Than Unconditional Ones?2022 · 5 citesDirectly Denoising Diffusion Models2024 · 2 citesOne-step Diffusion with Distribution Matching Distillation2023 · 1 citesACT-Diffusion: Efficient Adversarial Consistency Training for One-step Diffusion Models2023Posterior sampling via Langevin dynamics based on generative priors2024