Tag
Flash-BoN improves text-to-image generation efficiency by generating cheap draft candidates via timestep truncation, layer skipping, and activation proxies, then using multi-stage verification to select the best draft for full refinement, outperforming baselines under fixed wall-clock budgets.
NVIDIA released Nemotron-TwoTower-30B-A3B-Base-BF16, a diffusion-based language model that uses block-wise autoregressive diffusion to generate text by iterative denoising of token blocks, achieving 2.42× the generation throughput of the autoregressive baseline while retaining 98.7% of benchmark quality.
Dustin introduces a sparse verification framework for speculative decoding that leverages draft model signals and sparse attention head scoring to overcome the KV cache verification bottleneck, achieving up to 27.85x speedup in self-attention and 9.17x end-to-end decoding speedup on long-context tasks with negligible accuracy loss.
This paper introduces latent spatial memory for video world models, storing 3D scene information directly in diffusion latent space to avoid costly pixel-space reconstruction. The proposed Mirage framework achieves up to 10.57x faster generation and 55x memory reduction while achieving state-of-the-art performance on WorldScore and RealEstate10K.
SwiftI2V is a new efficient framework for high-resolution image-to-video generation that uses conditional segment-wise generation to achieve 2K synthesis with significantly reduced computational costs. It enables practical generation on single consumer or datacenter GPUs while maintaining input fidelity.
L2P proposes an efficient transfer paradigm that leverages pre-trained latent diffusion models to build pixel-space diffusion models, enabling high-quality generation with minimal computational overhead and data requirements, and supporting native 4K resolution.
NVlabs/Sana is an efficiency-oriented open-source codebase for high-resolution image and video generation, including multiple model variants and training/inference pipelines.