EDN

Why memory swizzling is hidden tax on AI compute

Memory swizzling is the quiet tax that every hierarchical-memory accelerator pays. It is fundamental to how GPUs, TPUs, NPUs, ...

Skip Subscriptions, Set up Fast Local AI for Coding, Study, and Brainstorming

Learn how to run local AI models with LM Studio's user, power user, and developer modes, keeping data private and saving monthly fees.
IEEE

Quantization via Distillation and Contrastive Learning

Abstract: Quantization is a critical technique employed across various research fields for compressing deep neural networks (DNNs) to facilitate deployment within resource-limited environments. This ...
IEEE

SearchQ: Search-Based Fine-Grained Quantization for Data-Free Model Compression

Abstract: The huge memory and computing costs of deep neural networks (DNNs) greatly hinder their deployment on resource-constrained devices with high efficiency. Quantization has emerged as an ...
GitHub

PTQ4SAM: Post-Training Quantization for Segment Anything (CVPR 2024)

Segment Anything Model (SAM) has achieved impressive performance in many computer vision tasks. However, as a large-scale model, the immense memory and computation costs hinder its practical ...
GitHub

APHQ-ViT: Post-Training Quantization with Average Perturbation Hessian Based Reconstruction for Vision Transformers

This repository contains the official PyTorch implementation for the CVPR 2025 paper "APHQ-ViT: Post-Training Quantization with Average Perturbation Hessian Based Reconstruction for Vision ...