Generative Image Models Comparison Hub
3 papers - avg viability 7.3
Top Papers
- DINO-SAE: DINO Spherical Autoencoder for High-Fidelity Image Reconstruction and Generation(8.0)
DINO-SAE is a high-fidelity image reconstruction and generation tool using hyperspherical model alignment.
- How Long Can Unified Multimodal Models Generate Images Reliably? Taming Long-Horizon Interleaved Image Generation via Context Curation(7.0)
UniLongGen enhances the reliability of long-form image generation by dynamically curating the model's memory, leading to improved fidelity and consistency.
- Variational Flow Maps: Make Some Noise for One-Step Conditional Generation(7.0)
Variational Flow Maps enable fast, high-quality conditional image generation by learning to adapt noise distributions for one-step sampling, offering a speed advantage over iterative diffusion models.