Efficient Hybrid Distillation: Synergizing Score and Adversarial Objectives for One-Step Diffusion
cs.CV
Submitted: 2026-11-02
Updated: 2026-11-02
Code: https://github.com/pf0607/EASD
Terminology
Sources
- Towards Principled Methods for Training Generative Adversarial Networks
- Large Scale GAN Training for High Fidelity Natural Image Synthesis
- Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer
- pi-Flow: Policy-Based Few-Step Generation via Imitation Distillation
- One Step Diffusion via Shortcut Models
- Consistency Models Made Easy
- Mean Flows for One-step Generative Modeling
- Improved Mean Flows: On the Challenges of Fastforward Generative Models
- Classifier-Free Diffusion Guidance
- Scalable Adaptive Computation for Iterative Generation
- Score-of-Mixture Training: Training One-Step Generative Models Made Simple via Score Estimation of Mixture Distributions
- Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion
- FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space
- Decoupled MeanFlow: Turning Flow Models into Flow Maps for Accelerated Sampling
- Truncated Consistency Models
- SDXL-Lightning: Progressive Adversarial Diffusion Distillation
- Pseudo Numerical Methods for Diffusion Models on Manifolds
- Learning to Integrate Diffusion ODEs by Averaging the Derivatives
- Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
- InstaFlow: One Step is Enough for High-Quality Diffusion-Based Text-to-Image Generation
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs