The paper investigates post‑training quantization of transformer attention blocks by optimizing a joint loss over the Q, K, V projections rather than individual weight matrices. Using this joint attention‑based objective (JAB), the authors achieve significant compression on Mistral‑7B, recovering 77‑90% of the performance gap at 3 bits, but the method fails when MLP layers are included. A role‑aware offset rule that ignores sensitivity estimates outperforms JAB on GPT‑2 and full Mistral‑7B, demonstrating that the matrix a weight belongs to is more critical than sensitivity metrics.
The study evaluates the portability of INT8 post‑training quantization across seven hardware platforms, including CPUs, GPUs, and vendor NPUs, by keeping the ONNX model and quantization scales constant. It finds that INT8 performance and output consistency vary significantly: CPU dot‑product instructions determine speedup, identical INT8 outputs only occur when integer kernels match, and vendor NPUs require their own quantization pipelines. The authors also show that edge‑NPU latency is dominated by data transfer rather than compute and provide scripts and reports for reproducibility.
By Yuyeong Shin
arXiv:2607. 17240v1 Announce Type: new Abstract: When does a committed intermediate stage in an LLM reasoning pipeline earn its cost?
By Honglin Li (ShanghaiTech University)
arXiv:2608.30564v1 Announce Type: cross
Abstract: Mixed-precision quantization (MPQ) assigns a different bitwidth to each linear layer of a large language model (LLM) to minimize the quantization-ind...
By Deokjae Lee, Sihun Chu, Hyun Oh Song
arXiv:2608.23601v1 Announce Type: cross
Abstract: EDA flow parameter tuning is critical for quality-of-results~(QoR), yet the parameter space is large, tightly coupled, and full evaluations are prohi...
By Kunlong Li, Shangshang Yao, Su Zheng, Lingli Wang
arXiv:2605. 24011v2 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models exhibit remarkable action generation for embodied intelligence, but their heavy compute make deployment on edge platforms impractical.
By Arash Akbari, Arman Akbari, Masih Eskandar, Qitao Tan, Yixiao Chen, Jingwu Luo, Bertha Pangaribuan, Liyun Zhang, Jennifer Dy, Geng Yuan, Xue Lin, Gaowen Liu, Stratis Ioannidis, Yanzhi Wang
Dynamic Expert Quantization (DynaExq) is a runtime-aware mixed-precision serving system designed for single‑GPU Mixture‑of‑Experts (MoE) inference under a hard high‑bandwidth memory (HBM) envelope. It treats the problem as an online, budget‑constrained precision allocation task, keeping the most frequently used experts at higher precision while relegating the rest to low‑precision fallbacks. By estimating expert hotness from router traces and asynchronously promoting or demoting experts, DynaExq maintains a fully materialized expert set during the forward pass, improving accuracy and throughput compared to static post‑training quantization and offloading/prefetch baselines.
whyItMatters":"DynaExq enables efficient deployment of large MoE models on memory‑limited GPUs by dynamically allocating precision based on runtime expert usage, thereby reducing memory footprint and latency while boosting accuracy and throughput."
By Kexin Chu, Dawei Xiang, Zixu Shen, Yiwei Yang, Zecheng Liu, Wei Zhang
arXiv:2606. 11387v1 Announce Type: cross Abstract: Short pretraining runs can reduce experimental cost, but they can also over-promote configurations that only look strong at tiny budgets.
By Felipe Chavarro Polania
arXiv:2608. 07911v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models have outgrown accelerator memory, and offloading expert weights to host memory is now standard.
By Yu Zhang
arXiv:2605. 14672v2 Announce Type: replace Abstract: Estimating an $N \times N$ quantum kernel from circuit fidelities requires $\Theta(N^2 S)$ measurement shots, the dominant bottleneck for deployment on near-term hardware.
By Jian Xu, Chao Li, Delu Zeng, John Paisley, Qibin Zhao
arXiv:2607. 05876v1 Announce Type: cross Abstract: LLM serving optimization typically benchmarks many configurations and reaches for heavy profilers when latency targets are missed.
By Yihua Liu
arXiv:2608.23816v1 Announce Type: new
Abstract: Quantized fine-tuning (QLoRA) saves memory but not time. It dequantizes every 4-bit weight on the fly, so it trains more slowly than fp16 LoRA. We pres...
By Md Romyull Islam