Fast Walsh-Hadamard Transform
PulseAugur coverage of Fast Walsh-Hadamard Transform — every cluster mentioning Fast Walsh-Hadamard Transform across labs, papers, and developer communities, ranked by signal.
-
New RSLM method slashes ANN search embedding size to 1-4 bits
Researchers have introduced RSLM, a novel training-free vector quantization method designed to enhance Approximate Nearest Neighbor (ANN) search systems. This technique compresses embeddings to as few as 1-4 bits per di…
-
New 6-bit audio codec inspired by LLM quantization techniques
Researchers have developed AudioTQ, a novel 6-bit audio codec that operates directly in the time domain, bypassing traditional psychoacoustic modeling. Inspired by techniques used in large language model weight quantiza…
-
New ShearFuse-UNet model predicts wildfire spread efficiently
Researchers have developed ShearFuse-UNet, a novel deep learning model designed for predicting wildfire spread using satellite data. This model is notable for its lightweight architecture and computational efficiency, i…
-
llama.cpp adds CUDA FWHT for faster KV cache quantization
A pull request to the llama.cpp project introduces a CUDA implementation of the Fast Walsh-Hadamard Transform (FWHT). This optimization, developed by user am17an, aims to speed up operations when quantizing the key-valu…