MACaQ: Fast, Model Agnostic Calibrated Quantization for Low-Precision Weights and Activations

Lorenz K. Muller, Philippe Bich, Chiara Boretti, Hyun-Min Chang, Jiawei Zhuang, Lukas Cavigelli

MACaQ: Fast, Model Agnostic Calibrated Quantization for Low-Precision Weights and Activations: 30 upvotes on Hugging Face Daily Papers, #7 of 39 papers on 2026-11-11. Day-by-day upvote history.

Upvotes over time

It had 20 upvotes by its first day, 30 upvotes by the next day, 30 upvotes by day 3, 30 upvotes by day 7, and has 30 now. It ranked #25 of 133 papers in its week and #69 of 340 in its month. Its code repository has 203 GitHub stars.

Weight-activation quantization can reduce the memory and compute cost of neural-network inference, but recent methods often require expensive optimization and architecture-specific propagation of basis transforms across layers. We introduce MACaQ, an MXFP4 W4A4 post-training quantization method that treats every linear layer independently. MACaQ applies a per-channel scale and a block-Hadamard transform online to each layer's activations, while absorbing the corresponding inverse transforms into its weights. Retaining these transforms online avoids cross-layer dependencies and enables a simple streaming implementation of the quantization process, in which only one layer must reside in memory at a time. To improve over round-to-nearest quantization, we introduce Flip-Refinement, a Hessian-aware coordinate-descent procedure that greedily moves individual FP4 codes to adjacent representable values. Despite executing the activation transforms at inference time, MACaQ adds only 2.3% end-to-end overhead over pure MXFP4 on Qwen3-32B in vLLM. MACaQ reaches lower perplexity than DuQuant++ and MR-GPTQ, and already matches their best perplexity on Llama-3-8B in 11 times less quantization time. We further apply MACaQ without architecture-specific modifications to an image diffusion model, illustrating the generality of its layer-local design and show quantization of a 122B parameter model in 64 minutes, using only 32.5GB GPU RAM.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.