OrbitQuant: Efficient Quantization for Diffusion Transformers

Training-Free Single-Image Diffusion Model Achieves Fast, High-Quality Generation

This paper presents a training-free approach to single-image diffusion models. Instead of training a neural network on a single image (which

Tauformer: A Topological Transformer Architecture Using Laplacian-Derived Scalar Attention

The article discusses Tauformer, a novel topological transformer architecture that replaces traditional dot-product attention with a Laplaci

DiffusionBench: A Holistic Benchmark for Evaluating Diffusion Transformers

This is a GitHub repository announcement for DiffusionBench, a unified benchmark and codebase for evaluating and training Diffusion Transfor

TurboQuant: Compressing AI Vectors to 2-4 Bits Using Random Rotations

TurboQuant is a novel compression technique for AI vectors (KV caches, embeddings, attention keys) that compresses each coordinate to 2-4 bi

Systematic Study Shows Transformers Can Drop One or More QKV Projections Without Quality Loss

This research paper systematically evaluates whether Transformers need all three QKV (query, key, value) projections in attention mechanisms

Systematic Study Shows Transformers Can Drop One or More QKV Projections Without Quality Loss

This research paper systematically evaluates whether Transformers need all three QKV (query, key, value) projections in attention mechanisms

Tiny-diffusion: A minimal implementation of probabilistic diffusion models