Related papers: ELT: Elastic Looped Transformers for Visual Genera…

Diffusion transformers (DiTs) achieve high generative quality but lock FLOPs to image resolution, limiting principled latency-quality trade-offs, and allocate computation uniformly across input spatial tokens, wasting resource allocation to…

Recent advances in distilling expensive diffusion models into efficient few-step generators show significant promise. However, these methods typically demand substantial computational resources and extended training periods, limiting…

Inverse lithography (ILT) is critical for modern semiconductor manufacturing but suffers from highly non-convex objectives that often trap optimization in poor local minima. Generative AI has been explored to warm-start ILT, yet most…

Diffusion Transformers (DiT) have demonstrated remarkable generative capabilities but remain highly computationally expensive. Previous acceleration methods, such as pruning and distillation, typically rely on a fixed computational…

While recent advancements in multimodal language models have enabled image generation from expressive multi-image instructions, existing methods struggle to maintain performance under complex interleaved instructions. This limitation stems…

While diffusion models can learn complex distributions, sampling requires a computationally expensive iterative process. Existing distillation methods enable efficient sampling, but have notable limitations, such as performance degradation…

Scaling video diffusion transformers is fundamentally bottlenecked by two compounding costs: the expensive quadratic complexity of attention per step, and the iterative sampling steps. In this work, we propose EFlow, an efficient few-step…

Recent advances in multimodal foundation models unifying image understanding and generation have opened exciting avenues for tackling a wide range of vision-language tasks within a single framework. Despite progress, existing unified models…

Although more layers and more parameters generally improve the accuracy of the models, such big models generally have high computational complexity and require big memory, which exceed the capacity of small devices for inference and incurs…

Parameter-efficient transfer learning (PETL) has shown great potential in adapting a vision transformer (ViT) pre-trained on large-scale datasets to various downstream tasks. Existing studies primarily focus on minimizing the number of…

With the rapid development of computer vision, Vision Transformers (ViTs) offer the tantalising prospect of unified information processing across visual and textual domains due to the lack of inherent inductive biases in ViTs. ViTs require…

Dataset distillation (DD) aims to synthesize compact training sets that enable models to achieve high accuracy with significantly fewer samples. Recent diffusion-based DD methods commonly introduce semantic guidance through late-stage…

Latent Diffusion Models (LDMs) enable high-quality image synthesis while avoiding excessive compute demands by training a diffusion model in a compressed lower-dimensional latent space. Here, we apply the LDM paradigm to high-resolution…

Most meta-learning approaches assume the existence of a very large set of labeled data available for episodic meta-learning of base knowledge. This contrasts with the more realistic continual learning paradigm in which data arrives…

The Efficient Adaptive Transformer (EAT) framework unifies three adaptive efficiency techniques - progressive token pruning, sparse attention, and dynamic early exiting - into a single, reproducible architecture for input-adaptive…

High-resolution image synthesis remains a core challenge in generative modeling, particularly in balancing computational efficiency with the preservation of fine-grained visual detail. We present Latent Wavelet Diffusion (LWD), a…

As neural networks are increasingly being applied to real-world applications, mechanisms to address distributional shift and sequential task learning without forgetting are critical. Methods incorporating network expansion have shown…

Efficient training strategies for large-scale diffusion models have recently emphasized the importance of improving discriminative feature representations in these models. A central line of work in this direction is representation alignment…

Latent variable generative models have emerged as powerful tools for generative tasks including image and video synthesis. These models are enabled by pretrained autoencoders that map high resolution data into a compressed lower dimensional…

The rapid development of generative models has significantly advanced image and video applications. Among these, video creation, aimed at generating videos under various conditions, has gained substantial attention. However, existing video…