Quantization-Aware Training

Also known as QAT

Overview

Quantization-aware training (QAT) simulates low-precision arithmetic during training so models can learn to withstand rounding and clipping errors at deployment. Common approaches use straight-through estimators, learned or mixed-precision quantizers, progressive bit reduction, and teacher-guided distillation. The methods are applied to settings from ImageNet vision models and edge devices to large language models, where evaluations include accuracy, perplexity, reasoning, and real-device speed. A central challenge is balancing accuracy against bit width and hardware cost: results depend on the numerical format and deployment path, and low-precision training can become unstable or diverge from inference behavior.

Latest papers

94 papers11 papers in the last four weeks, down 39% on the four weeks before. 0.1% of all new papers.

May 5, 2026

Topology-Constrained Quantized nnUNet for Efficient and Anatomically Accurate 3D Tooth Segmentation

We propose a topology-constrained quantized nnUNet framework for efficient and anatomically accurate 3D tooth segmentation, addressing the challenges of spatial distortion introduced by quantization in deep learning models. The proposed method integrates a novel tooth-specific topological loss into quantization-aware training, preserving critical anatomical structures such as tooth count, adjacency relationships, and cavity integrity while maintaining computational efficiency. The system employs an 8-bit quantized nnUNet backbone, where weights and activations are dynamically calibrated to minimize precision loss during inference. Furthermore, the topological loss combines connected-component analysis, adjacency consistency, and hole detection penalties, ensuring anatomical fidelity without modifying the underlying network architecture. The joint optimization objective harmonizes cross-entropy loss, quantization regularization, and topological constraints, enabling end-to-end training with gradient approximations for persistent homology terms. Experiments demonstrate that our approach significantly reduces topological errors compared to conventional quantized models, achieving clinically plausible segmentations on dental CBCT scans. The method retains the hardware efficiency of integer-only inference, making it suitable for deployment in resource-constrained clinical environments. This work bridges the gap between computational efficiency and anatomical precision in medical image segmentation, offering a practical solution for real-world dental applications.
May 4, 2026

DurableUn: Quantization-Induced Recovery Attacks in Machine Unlearning

Machine unlearning aims to remove specified training data to satisfy privacy regulations such as GDPR. However, existing evaluations assume identical precision at unlearning and deployment, overlooking that production LLMs are deployed at low-bit precision. We show that INT4 quantization systematically restores forgotten content even when models pass compliance audits at bfloat16 (BF16), we term this the quantization recovery attack (QRA). We conduct the first systematic study of unlearning robustness under adapter-space INT4 quantization in the NF4+LoRA regime, evaluating seven methods on LLaMA-3-8B-Instruct across TOFU, MUSE-News, and WikiBio-WPU. INT8 is benign; INT4 induces recovery of up to 22x, worsening with dataset difficulty. We identify the FA-RA-Q-INT4 trilemma: no method simultaneously achieves strong forgetting, high utility, and quantization robustness. A dense Pareto sweep reveals a sharp phase transition once robustness is achieved, retaining accuracy collapses regardless of further tuning. To address this, we propose DURABLEUN-SAF (Sharpness-Aware Forgetting), a quantization-aware objective using Straight-Through Estimator gradients through INT4 rounding. DURABLEUN-SAF is the only method to achieve a stable empirical (0.047, {BF16, INT8, INT4})- durability certificate: Q-INT4= 0.043 +- 0.002, cert rate= 3/3, versus SalUn's cert rate= 1/3 at its own published hyperparameters. We call for Q-INT4 to be adopted as a standard evaluation metric alongside FA and RA.
Apr 25, 2026

Efficient VQ-QAT and Mixed Vector/Linear quantized Neural Networks

In this work, we developed and tested 3 techniques for vector quantization (VQ) based model weight compression. To mitigate codebook collapse and enable end-to-end training, we adopted cosine similarity-based assignment. Building on ideas from attention-based formulations in Differentiable K-Means (DKM), we further improved this approach by using cosine similarity for assignment combined with top-1 sampling and a straight-through estimator, thereby eliminating the need for weighted-average reconstruction. Finally, we investigated the use of differentiable neural architecture search (NAS) to adaptively select layer-wise quantization configurations, further optimizing the compression process. Although our method does not consistently outperform existing approaches across all quantization levels, it provides useful insights into the design trade-offs and behaviors of VQ-based model compression methods.
Apr 22, 2026

Efficient INT8 Single-Image Super-Resolution via Deployment-Aware Quantization and Teacher-Guided Training

Efficient single-image super-resolution (SISR) requires balancing reconstruction fidelity, model compactness, and robustness under low-bit deployment, which is especially challenging for x3 SR. We present a deployment-oriented quantized SISR framework based on an extract-refine-upsample design. The student performs most computation in the low-resolution space and uses a lightweight re-parameterizable backbone with PixelShuffle reconstruction, yielding a compact inference graph. To improve quality without significantly increasing complexity, we adopt a three-stage training pipeline: Stage 1 learns a basic reconstruction mapping with spatial supervision; Stage 2 refines fidelity using Charbonnier loss, DCT-domain supervision, and confidence-weighted output-level distillation from a Mamba-based teacher; and Stage 3 applies quantization-aware training directly on the fused deploy graph. We further use weight clipping and BatchNorm recalibration to improve quantization stability. On the MAI 2026 Quantized 4K Image Super-Resolution Challenge test set, our final AIO MAI submission achieves 29.79 dB PSNR and 0.8634 SSIM, obtaining a final score of 1.8 under the target mobile INT8 deployment setting. Ablation on Stage 3 optimization shows that teacher-guided supervision improves the dynamic INT8 TFLite reconstruction from 29.91 dB/0.853 to 30.0003 dB/0.856, while the fixed-shape deployable INT8 TFLite artifact attains 30.006 dB/0.857.
Apr 11, 2026

Device-Conditioned Neural Architecture Search for Efficient Robotic Manipulation

The growing complexity of visuomotor policies poses significant challenges for deployment with heterogeneous robotic hardware constraints. However, most existing model-efficient approaches for robotic manipulation are device- and model-specific, lack generalizability, and require time-consuming per-device optimization during the adaptation process. In this work, we propose a unified framework named Device-Conditioned Quantization-For-All (DC-QFA) which amortizes deployment effort with the device-conditioned quantization-aware training and hardware-constrained architecture search. Specifically, we introduce a single supernet that spans a rich design space over network architectures and mixed-precision bit-widths. It is optimized with latency- and memory-aware regularization, guided by per-device lookup tables. With this supernet, for each target platform, we can perform a once-for-all lightweight search to select an optimal subnet without any per-device re-optimization, which enables more generalizable deployment across heterogeneous hardware, and substantially reduces deployment time. To improve long-horizon stability under low precision, we further introduce multi-step on-policy distillation to mitigate error accumulation during closed-loop execution. Extensive experiments on three representative policy backbones, such as DiffusionPolicy-T, MDT-V, and OpenVLA-OFT, demonstrate that our DC-QFA achieves 2-3×2\text{-}3\times acceleration on edge devices, consumer-grade GPUs, and cloud platforms, with negligible performance drop in task success. Real-world evaluations on an Inovo robot equipped with a force/torque sensor further validates that our low-bit DC-QFA policies maintain stable, contact-rich manipulation even under severe quantization.
Apr 10, 2026

EdgeRazor: A Lightweight Framework for Large Language Models via Mixed-Precision Quantization-Aware Distillation

Quantization has emerged as a mainstream approach for deploying Large Language Models (LLMs) on resource-constrained devices, yet compressing precision below 4-bit typically causes severe performance degradation or prohibitive retraining costs. In this paper, we propose EdgeRazor, a lightweight framework for LLMs via Mixed-Precision Quantization-Aware Distillation. It contains three modules: Structural Quantization with Mixed Precision for fine-grained control of bit-widths, Layer-Adaptive Feature Distillation that dynamically selects the most informative features for alignment, and Entropy-Aware KL Divergence for forward-reverse balance on both human-annotated and distilled datasets. Evaluations conducted on MobileLLM and Qwen families show that under weight-activation quantization, the 1.88-bit Qwen3-0.6B-EdgeRazor outperforms the state-of-the-art 2-bit baselines by 11.27 and surpasses the strongest 3-bit baselines by 4.38, while the quantized MobileLLM-350M-EdgeRazor requires a training budget 4-10×\times lower than the leading quantization-aware training method. In terms of efficiency, EdgeRazor achieves higher compression ratios at all bit-widths, and the 1.58-bit Qwen3-0.6B-EdgeRazor reduces storage from 1.11 GB to 0.19 GB while accelerating decoding by 15.16×\times over the 16-bit baseline. These results empirically validate the effectiveness and efficiency of EdgeRazor. The codes can be accessed from GitHub and Huggingface.
Apr 9, 2026

Bit-by-Bit: Progressive QAT Strategy with Outlier Channel Splitting for Stable Low-Bit LLMs

Training LLMs at ultra-low precision remains a formidable challenge. Direct low-bit QAT often suffers from convergence instability and substantial training costs, exacerbated by quantization noise from heavy-tailed outlier channels and error accumulation across layers. To address these issues, we present Bit-by-Bit, a progressive QAT framework with outlier channel splitting. Our approach integrates three key components: (1) block-wise progressive training that reduces precision stage by stage, ensuring stable initialization for low-bit optimization; (2) nested structure of integer quantization grids to enable a "train once, deploy any precision" paradigm, allowing a single model to support multiple bit-widths without retraining; (3) rounding-aware outlier channel splitting, which mitigates quantization error while acting as an identity transform that preserves the quantized outputs. Furthermore, we follow microscaling groups with E4M3 scales, capturing dynamic activation ranges in alignment with OCP/NVIDIA standards. To address the lack of efficient 2-bit kernels, we developed custom operators for both W2A2 and W2A16 configurations, achieving up to 11×\times speedup over BF16. Under W2A2 settings, Bit-by-Bit significantly outperforms baselines like BitDistiller and EfficientQAT on both Llama2/3, achieving a loss of only 2.25 WikiText2 PPL compared to full-precision models.
Apr 9, 2026

QaRL: Rollout-Aligned Quantization-Aware RL for Fast and Stable Training under Training--Inference Mismatch

Large language model (LLM) reinforcement learning (RL) pipelines are often bottlenecked by rollout generation, making end-to-end training slow. Recent work mitigates this by running rollouts with quantization to accelerate decoding, which is the most expensive stage of the RL loop. However, these setups destabilize optimization by amplifying the training-inference gap: rollouts are operated at low precision, while learning updates are computed at full precision. To address this challenge, we propose QaRL (Rollout Alignment Quantization-Aware RL), which aligns training-side forward with the quantized rollout to minimize mismatch. We further identify a failure mode in quantized rollouts: long-form responses tend to produce repetitive, garbled tokens (error tokens). To mitigate these problems, we introduce TBPO (Trust-Band Policy Optimization), a sequence-level objective with dual clipping for negative samples, aimed at keeping updates within the trust region. On Qwen3-30B-A3B MoE for math problems, QaRL outperforms quantized-rollout training by +5.5 while improving stability and preserving low-bit throughput benefits.
Apr 5, 2026

BWTA: Accurate and Efficient Binarized Transformer by Algorithm-Hardware Co-design

Ultra low-bit quantization brings substantial efficiency for Transformer-based models, but the accuracy degradation and limited GPU support hinder its wide usage. In this paper, we analyze zero-point distortion in binarization and propose a Binary Weights & Ternary Activations (BWTA) quantization scheme, which projects tiny values to zero and preserves the accuracy of extremely low-bit models. For training, we propose Smooth Multi-Stage Quantization, combining a Levelwise Degradation Strategy and a Magnitude-Alignment Projection Factor to enable stable and fast convergence. For inference, we develop a BWTA MatMul CUDA kernel with instruction-level parallel bit-packing and comprehensive binary/ternary MatMul implementations for both linear and attention operators, allowing seamless integration across Transformer architectures. Experiments show that BWTA approaches full-precision performance for BERT, with an average 3.5% drop on GLUE and less than 2% drop on five tasks, and achieves comparable perplexity and accuracy for LLMs. In efficiency, it delivers 16 to 24 times kernel-level speedup over FP16 on NVIDIA GPUs, and 216 to 330 tokens/s end-to-end prefill speedup with lower memory footprint on LLMs. As an algorithm-hardware co-design, BWTA demonstrates practical, low-latency ultra-low-bit inference without sacrificing model quality.
Apr 2, 2026

AdaHOP: Fast and Accurate Low-Precision Training via Outlier-Pattern-Aware Rotation

Hadamard transforms have become a key tool for stabilizing low-precision training, but existing methods apply them uniformly across tensors and computation paths. We show that this one-size-fits-all strategy is inherently limited: Hadamard smoothing reduces quantization error only when its direction is properly aligned with the operand's outlier structure. Through a systematic study of weights, activations, and gradients in LLM training, we identify three stable outlier patterns, Row-wise, Column-wise, and None, and show that each outlier pattern pair in matrix multiplication requires a distinct transform or outlier-handling strategy. We propose AdaHOP, Adaptive Hadamard transform with Outlier-Pattern-aware strategy, which applies Inner Hadamard Transform (IHT) when inner-dimension mixing properly suppresses the operands' outliers, and selectively applies Outlier Extraction (OE) that extracts dominant outlier rows or columns into a high-precision path when it does not. With fused, hardware-aware Triton kernels, AdaHOP enables training from scratch at MXFP4 precision with BF16-level quality, while achieving up to 3.6X memory compression, 1.46X end-to-end training speedup over BF16.
Apr 1, 2026

AdaLoRA-QAT: Adaptive Low-Rank and Quantization-Aware Segmentation

Chest X-ray (CXR) segmentation is an important step in computer-aided diagnosis, yet deploying large foundation models in clinical settings remains challenging due to computational constraints. We propose AdaLoRA-QAT, a two-stage fine-tuning framework that combines adaptive low-rank encoder adaptation with full quantization-aware training. Adaptive rank allocation improves parameter efficiency, while selective mixed-precision INT8 quantization preserves structural fidelity crucial for clinical reliability. Evaluated across large-scale CXR datasets, AdaLoRA-QAT achieves 95.6% Dice, matching full-precision SAM decoder fine-tuning while reducing trainable parameters by 16.6\times and yielding 2.24\times model compression. A Wilcoxon signed-rank test confirms that quantization does not significantly degrade segmentation accuracy. These results demonstrate that AdaLoRA-QAT effectively balances accuracy, efficiency, and structural trust-worthiness, enabling compact and deployable foundation models for medical image segmentation. Code and pretrained models are available at: https://prantik-pdeb.github.io/adaloraqat.github.io/
Apr 1, 2026

MF-QAT: Multi-Format Quantization-Aware Training for Elastic Inference

Quantization-aware training (QAT) is typically performed for a single target numeric format, while practical deployments often need to choose numerical precision at inference time based on hardware support or runtime constraints. We study multi-format QAT, where a single model is trained to be robust across multiple quantization formats. We find that multi-format QAT can match single-format QAT at each target precision, yielding one model that performs well overall across different formats, even formats that were not seen during training. To enable practical deployment, we propose the Slice-and-Scale conversion procedure for both MXINT and MXFP that converts a high-precision representation into lower-precision formats without re-training. Building on this, we introduce a pipeline that (i) trains a model with multi-format QAT, (ii) stores a single anchor format checkpoint (MXINT8/MXFP8), and (iii) allows on-the-fly conversion to lower MXINT or MXFP formats at runtime with negligible-or no-additional accuracy degradation. Together, these components provide a practical path to elastic precision scaling and allow selecting the runtime format at inference time across diverse deployment targets.
Mar 31, 2026

Quantization with Unified Adaptive Distillation to enable multi-LoRA based one-for-all Generative Vision Models on edge

Generative Artificial Intelligence (GenAI) features such as image editing, object removal, and prompt-guided image transformation are increasingly integrated into mobile applications. However, deploying Large Vision Models (LVMs) for such tasks on resource-constrained devices remains challenging due to their high memory and compute requirements. While Low-Rank Adapters (LoRAs) enable parameter-efficient task adaptation, existing Mobile deployment pipelines typically compile separate model binaries for each LoRA + a copy of the foundation model, resulting in redundant storage and increased runtime overhead. In this work, we present a unified framework for enabling multi-task GenAI inference on edge devices using a single shared model. Our key idea is to treat LoRA weights as runtime inputs rather than embedding them into the compiled model graph, allowing dynamic task switching at runtime without recompilation. Then, to support efficient on-device execution, we introduce QUAD (Quantization with Unified Adaptive Distillation), a quantizationaware training strategy that aligns multiple LoRA adapters under a shared quantization profile. We implement the proposed system with a lightweight runtime stack compatible with mobile NPUs and evaluate it across multiple chipsets. Experimental results demonstrate up to 6x and 4x reduction in memory footprint and latency improvements, respectively, while maintaining high visual quality across multiple GenAI tasks.
Mar 19, 2026

SynQ: Accurate Zero-shot Quantization by Synthesis-aware Fine-tuning

How can we accurately quantize a pre-trained model without any data? Quantization algorithms are widely used for deploying neural networks on resource-constrained edge devices. Zero-shot Quantization (ZSQ) addresses the crucial and practical scenario where training data are inaccessible for privacy or security reasons. However, three significant challenges hinder the performance of existing ZSQ methods: 1) noise in the synthetic dataset, 2) predictions based on off-target patterns, and the 3) misguidance by erroneous hard labels. In this paper, we propose SynQ (Synthesis-aware Fine-tuning for Zero-shot Quantization), a carefully designed ZSQ framework to overcome the limitations of existing methods. SynQ minimizes the noise from the generated samples by exploiting a low-pass filter. Then, SynQ trains the quantized model to improve accuracy by aligning its class activation map with the pre-trained model. Furthermore, SynQ mitigates misguidance from the pre-trained model's error by leveraging only soft labels for difficult samples. Extensive experiments show that SynQ provides the state-of-the-art accuracy, over existing ZSQ methods.
Mar 14, 2026

True 4-Bit Quantized Convolutional Neural Network Training on CPU: Achieving Full-Precision Parity

Low-precision neural network training has emerged as a promising direction for reducing computational costs and democratizing access to deep learning research. However, existing 4-bit quantization methods either rely on expensive GPU infrastructure or suffer from significant accuracy degradation. In this work, we present a practical method for training convolutional neural networks at true 4-bit precision using standard PyTorch operations on commodity CPUs. We introduce a novel tanh-based soft weight clipping technique that, combined with symmetric quantization, dynamic per-layer scaling, and straight-through estimators, achieves stable convergence and competitive accuracy. Training a VGG-style architecture with 3.25 million parameters from scratch on CIFAR-10, our method achieves 92.34% test accuracy on Google Colab's free CPU tier -- matching full-precision baseline performance (92.5%) with only a 0.16% gap. We further validate on CIFAR-100, achieving 70.94% test accuracy across 100 classes with the same architecture and training procedure, demonstrating that 4-bit training from scratch generalizes to harder classification tasks. Both experiments achieve 8x memory compression over FP32 while maintaining exactly 15 unique weight values per layer throughout training. We additionally validate hardware independence by demonstrating rapid convergence on a consumer mobile device (OnePlus 9R), achieving 83.16% accuracy in only 6 epochs. To the best of our knowledge, no prior work has demonstrated 4-bit quantization-aware training achieving full-precision parity on standard CPU hardware without specialized kernels or post-training quantization.
Mar 11, 2026

The Curse and Blessing of Mean Bias in FP4-Quantized LLM Training

FP4 training promises substantial memory and compute savings for large language models, but remains fragile because blockwise quantization is dictated by extreme activation magnitudes, which inflate dynamic range and compress long-tail signals. We identify a counterintuitive source of this failure: dominant activation outliers are not merely arbitrary sparse events, but are largely induced by a coherent rank-one mean bias, whose direction aligns with the leading anisotropic spectral component. This mean component strengthens during training, is amplified and reshaped by attention and FFN operators, and increasingly dominates top activation magnitudes. Crucially, this discovery reveals that a seemingly complex outlier-suppression problem admits a truly simple solution: isolate the coherent mean before quantization. We therefore propose Averis, a mean-residual splitting quantization method that separates the mean component using only reductions and elementwise subtractions before FP4 quantization. Across Qwen3 0.6B Dense trained on 100B tokens and Qwen3 7B A1.5B MoE trained on 50B tokens, Averis enables robust W4A4G4 FP4 training, reducing BF16 loss gaps to 1.19%/0.81% versus 2.05%/1.10% for NVIDIA's recently released Hadamard-based outlier-smoothing method, while limiting downstream gaps to 0.89/0.71 points. With only 2.20% end-to-end overhead over vanilla NVFP4, about 30% of NVIDIA's Hadamard-based design, Averis provides a hardware-efficient path to stable low-bit LLM training. Complementary to Hadamard, Averis further reduces the Qwen3-0.6B loss and downstream gaps to 0.94% and 0.73 points when combined. Code is available at: https://anonymous.4open.science/r/averis-504D.
Mar 10, 2026

Decoder-Free Distillation for Quantized Image Restoration

Quantization-Aware Training (QAT), combined with Knowledge Distillation (KD), holds immense promise for compressing models for edge deployment. However, joint optimization for precision-sensitive image restoration (IR) to recover visual quality from degraded images remains largely underexplored. Directly adapting QAT-KD to low-level vision reveals three critical bottlenecks: teacher-student capacity mismatch, spatial error amplification during decoder distillation, and an optimization "tug-of-war" between reconstruction and distillation losses caused by quantization noise. To tackle these, we introduce Quantization-aware Distilled Restoration (QDR), a framework for edge-deployed IR. QDR eliminates capacity mismatch via FP32 self-distillation and prevents error amplification through Decoder-Free Distillation (DFD), which corrects quantization errors strictly at the network bottleneck. To stabilize the optimization tug-of-war, we propose a Learnable Magnitude Reweighting (LMR) that dynamically balances competing gradients. Finally, we design an Edge-Friendly Model (EFM) featuring a lightweight Learnable Degradation Gating (LDG) to dynamically modulate spatial degradation localization. Extensive experiments across four IR tasks demonstrate that our Int8 model recovers 96.5% of FP32 performance, achieves 442 frames per second (FPS) on an NVIDIA Jetson Orin, and boosts downstream object detection by 16.3 mAP
Mar 6, 2026

QATMA: Quantization-Aware Training with Multimodal Alignment for Open-Vocabulary Object Detection

Quantizing open-vocabulary object detection (OVOD) models reduces their memory and computational costs, but extremely low-bit quantization severely degrades both cross-modal (region-text) and intra-modal (region-region) alignments. This multimodal degradation is a unique challenge that prior quantization methods for closed-vocabulary detectors fail to resolve. To overcome this, we propose Quantization-Aware Training with Multimodal Alignment (QATMA), the first multimodal-aware and architecture-agnostic QAT framework tailored for OVOD. QATMA integrates two key components: (i) Curriculum QAT, which partitions the detector by functional roles and progressively expands the quantization scope to suppress error accumulation and ensure stable optimization; and (ii) Text-anchored Pairwise Similarity Distillation, which transfers both region-text and region-region alignments from a full-precision teacher model via pairwise cosine similarities in the joint embedding space. Experimental results on LVIS and COCO zero-shot benchmarks demonstrate that QATMA significantly outperforms existing QAT baselines under extremely low-bit settings, achieving gains of up to 4.3 and 7.6 AP, respectively.
Mar 5, 2026

Trainable Bitwise Soft Quantization for Input Feature Compression

The growing demand for machine learning applications in the context of the Internet of Things calls for new approaches to optimize the use of limited compute and memory resources. Despite significant progress that has been made w.r.t. reducing model sizes and improving efficiency, many applications still require remote servers to provide the required resources. However, such approaches rely on transmitting data from edge devices to remote servers, which may not always be feasible due to bandwidth, latency, or energy constraints. We propose a task-specific, trainable feature quantization layer that compresses the input features of a neural network. This can significantly reduce the amount of data that needs to be transferred from the device to a remote server. In particular, the layer allows each input feature to be quantized to a user-defined number of bits, enabling a simple on-device compression at the time of data collection. The layer is designed to approximate step functions with sigmoids, enabling trainable quantization thresholds. By concatenating outputs from multiple sigmoids, introduced as bitwise soft quantization, it achieves trainable quantized values when integrated with a neural network. We compare our method to full-precision inference as well as to several quantization baselines. Experiments show that our approach outperforms standard quantization methods, while maintaining accuracy levels close to those of full-precision models. In particular, depending on the dataset, compression factors of 5×5\times to 16×16\times can be achieved compared to 3232-bit input without significant performance loss.
Mar 2, 2026

SageBwd: A Trainable Low-bit Attention

Low-bit attention, such as SageAttention, has emerged as an effective approach for accelerating model inference, but its applicability to training remains poorly understood. In prior work, we introduced SageBwd, a trainable INT8 attention that quantizes six of seven attention matrix multiplications while preserving fine-tuning performance. However, SageBwd exhibited a persistent performance gap to full-precision attention (FPA) during pre-training. In this work, we investigate why this gap occurs and demonstrate that SageBwd matches full-precision attention during pretraining. Through experiments and theoretical analysis, we reach a few important insights and conclusions: (i) QK-norm is necessary for stable training at large tokens per step, (ii) quantization errors primarily arise from the backward-pass score gradient dS, (iii) reducing tokens per step enables SageBwd to match FPA performance in pre-training, and (iv) K-smoothing remains essential for training stability, while Q-smoothing provides limited benefit during pre-training.
Mar 2, 2026

Boosting Entropy with Bell Box Quantization

Quantization-Aware Pre-Training (QAPT) is an effective technique to reduce the compute and memory overhead of Deep Neural Networks while improving their energy efficiency on edge devices. Existing QAPT methods produce models stored in compute-efficient data types (e.g. integers) that are not information theoretically optimal (ITO). On the other hand, existing ITO data types (e.g. Quantile/NormalFloat Quantization) are not compute-efficient. We propose BBQ, the first ITO quantization method that is also compute-efficient. BBQ builds on our key insight that since learning is domain-agnostic, the output of a quantizer does not need to reside in the same domain as its input. BBQ performs ITO quantization in its input domain, and returns its output in a compute-efficient domain where ITO data types are mapped to compute-efficient data types. Without sacrificing compute efficiency, BBQ outperforms prior SOTA QAPT methods by a perplexity reduction of up to 2 points for 4-bit models, up to 4 points for 3-bit models, up to 5 points for 2-bit models, and up to 18 points for 1-bit models. Code is available at https://github.com/1733116199/bbq.
Feb 26, 2026

FairQuant: Fairness-Aware Mixed-Precision Quantization for Medical Image Classification

Compressing neural networks by quantizing model parameters offers useful trade-off between performance and efficiency. Methods like quantization-aware training and post-training quantization strive to maintain the downstream performance of compressed models compared to the full precision models. However, these techniques do not explicitly consider the impact on algorithmic fairness. In this work, we study fairness-aware mixed-precision quantization schemes for medical image classification under explicit bit budgets. We introduce FairQuant, a framework that combines group-aware importance analysis, budgeted mixed-precision allocation, and a learnable Bit-Aware Quantization (BAQ) mode that jointly optimizes weights and per-unit bit allocations under bitrate and fairness regularization. We evaluate the method on Fitzpatrick17k and ISIC2019 across ResNet18/50, DeiT-Tiny, and TinyViT. Results show that FairQuant configurations with average precision near 4-6 bits recover much of the Uniform 8-bit accuracy while improving worst-group performance relative to Uniform 4- and 8-bit baselines, with comparable fairness metrics under shared budgets.
Feb 26, 2026

pQuant: Towards Effective Low-Bit Language Models via Decoupled Linear Quantization-Aware Training

Quantization-Aware Training from scratch has emerged as a promising approach for building efficient large language models (LLMs) with extremely low-bit weights (sub 2-bit), which can offer substantial advantages for edge deployment. However, existing methods still fail to achieve satisfactory accuracy and scalability. In this work, we identify a parameter democratization effect as a key bottleneck: the sensitivity of all parameters becomes homogenized, severely limiting expressivity. To address this, we propose pQuant, a method that decouples parameters by splitting linear layers into two specialized branches: a dominant 1-bit branch for efficient computation and a compact high-precision branch dedicated to preserving the most sensitive parameters. Through tailored feature scaling, we explicitly guide the model to allocate sensitive parameters to the high-precision branch. Furthermore, we extend this branch into multiple, sparsely-activated experts, enabling efficient capacity scaling. Extensive experiments indicate our pQuant achieves state-of-the-art performance in extremely low-bit quantization.
Feb 21, 2026

Rank-Aware Spectral Bounds on Attention Logits for Stable Low-Precision Training

Attention scores in transformers are bilinear forms Sij=xi⊤Mxj/dhS_{ij} = x_i^\top M x_j / \sqrt{d_h} whose maximum magnitude governs overflow risk in low-precision training. We derive a rank-aware concentration inequality: when the interaction matrix M=WQWK⊤M = W^Q W^{K\top} has rank r≪dr \ll d, tail probabilities for max⁡i,j∣Sij∣\max_{i,j}|S_{ij}| decay as exp⁡(−d2α2/(γr))\exp(-d^{2}α^{2}/(γr)) rather than exp⁡(−dα2)\exp(-dα^{2}), where γ>1γ> 1. For transformer attention where r=dhr = d_h, this yields 88 to 28×28\times tighter concentration than rank-agnostic bounds in modern architectures. We apply this result to FP8 training, deriving geometry-aware scale factors that provide principled overflow guarantees without observing activations. The method computes per-layer scales from the spectral norm ∥WQWK⊤∥2\|W^Q W^{K\top}\|_2 via implicit power iteration, includes a grouped query attention formulation that avoids key expansion, and remains compatible with fused attention kernels. Across GPT-2 XL to Llama-2-70B, geometry-aware scaling eliminates overflows in transient scenarios where delayed scaling fails, while achieving comparable downstream MMLU accuracy.
Feb 21, 2026

UFO: Unlocking Ultra-Efficient Quantized Private Inference with Protocol and Algorithm Co-Optimization

Private convolutional neural network (CNN) inference based on secure two-party computation (2PC) suffers from high communication and latency overhead, especially from convolution layers. In this paper, we propose UFO, a quantized 2PC inference framework that jointly optimizes the 2PC protocols and quantization algorithm. UFO features a novel 2PC protocol that systematically combines the efficient Winograd convolution algorithm with quantization to improve inference efficiency. However, we observe that naively combining quantization and Winograd convolution faces the following challenges: 1) From the inference perspective, Winograd transformations introduce extensive additions and require frequent bit width conversions to avoid inference overflow, leading to non-negligible communication overhead; 2) From the training perspective, Winograd transformations introduce weight outliers that make quantization-aware training (QAT) difficult, resulting in inferior model accuracy. To address these challenges, we co-optimize both protocol and algorithm. 1) At the protocol level, we propose a series of graph-level optimizations for 2PC inference to minimize the communication. 2) At the algorithm level, we develop a mixed-precision QAT algorithm based on layer sensitivity to optimize model accuracy given communication constraints. To accommodate the outliers, we further introduce a 2PC-friendly bit re-weighting algorithm to increase the representation range without explicitly increasing bit widths. With extensive experiments, UFO demonstrates 11.7x, 3.6x, and 6.3x communication reduction with 1.29%, 1.16%, and 1.29% higher accuracy compared to state-of-the-art frameworks SiRNN, COINN, and CoPriv, respectively.
Feb 17, 2026

1-Bit Wonder: Improving QAT Performance in the Low-Bit Regime through K-Means Quantization

Quantization-aware training (QAT) is an effective method to drastically reduce the memory footprint of LLMs while keeping performance degradation at an acceptable level. However, the optimal choice of quantization format and bit-width presents a challenge in practice. The full design space of quantization is not fully explored in the context of QAT, and the precise trade-off between quantization and downstream performance is poorly understood, as comparisons often rely solely on perplexity-based evaluations. In this work, we address these shortcomings with an empirical study of QAT in the low-bit regime. We show that k-means based weight quantization outperforms integer formats and can be implemented efficiently on standard hardware. Furthermore, we find that, under a fixed inference memory budget, the best performance on generative downstream tasks is achieved with 11-bit quantized weights.
Feb 3, 2026

LSGQuant: Layer-Sensitivity Guided Quantization for One-Step Diffusion Real-World Video Super-Resolution

One-Step Diffusion Models have demonstrated promising capability and fast inference in video super-resolution (VSR) for real-world. Nevertheless, the substantial model size and high computational cost of Diffusion Transformers (DiTs) limit downstream applications. While low-bit quantization is a common approach for model compression, the effectiveness of quantized models is challenged by the high dynamic range of input latent and diverse layer behaviors. To deal with these challenges, we introduce LSGQuant, a layer-sensitivity guided quantizing approach for one-step diffusion-based real-world VSR. Our method incorporates a Dynamic Range Adaptive Quantizer (DRAQ) to fit video token activations. Furthermore, we estimate layer sensitivity and implement a Variance-Oriented Layer Training Strategy (VOLTS) by analyzing layer-wise statistics in calibration. We also introduce Quantization-Aware Optimization (QAO) to jointly refine the quantized branch and a retained high-precision branch. Extensive experiments demonstrate that our method has nearly performance to origin model with full-precision and significantly exceeds existing quantization techniques. Code is available at: https://github.com/zhengchen1999/LSGQuant.
Jan 30, 2026

Gated Relational Alignment via Confidence-based Distillation for Efficient VLMs

Vision-Language Models (VLMs) achieve strong multimodal performance but are costly to deploy, and post-training quantization often causes significant accuracy loss. Despite its potential, quantization-aware training for VLMs remains underexplored. We propose GRACE, a framework unifying knowledge distillation and QAT under the Information Bottleneck principle: quantization constrains information capacity while distillation guides what to preserve within this budget. Treating the teacher as a proxy for task-relevant information, we introduce confidence-gated decoupled distillation to filter unreliable supervision, relational centered kernel alignment to transfer visual token structures, and an adaptive controller via Lagrangian relaxation to balance fidelity against capacity constraints. Across extensive benchmarks on LLaVA and Qwen families, our INT4 models consistently outperform FP16 baselines (e.g., LLaVA-1.5-7B: 70.1 vs. 66.8 on SQA; Qwen2-VL-2B: 76.9 vs. 72.6 on MMBench), nearly matching teacher performance. Using real INT4 kernel, we achieve 3×\times throughput with 54% memory reduction. This principled framework significantly outperforms existing quantization methods, making GRACE a compelling solution for resource-constrained deployment. Code and data are available at: https://github.com/ForeverBlue816/GRACE.
Nov 24, 2025

QuantKAN: A Unified Quantization Framework for Kolmogorov Arnold Networks

Kolmogorov--Arnold Networks (KANs) replace linear weights with spline-based functions, offering strong expressivity but posing challenges for low-precision deployment due to heterogeneous parameter distributions. We introduce QuantKAN, the first unified framework for quantization-aware training (QAT) and post-training quantization (PTQ) of KANs. The framework employs branch-aware quantizers for base and spline parameters and extends modern QAT and PTQ methods to spline-based layers across EfficientKAN, FastKAN, PyKAN, and KAGN. Experiments on MNIST, CIFAR-10/100, TinyImageNet, and ImageNet provide the first unified QAT/PTQ KAN benchmarks and show that DSQ is the most robust QAT method at aggressive low-bit settings, while GPTQ is the strongest PTQ method at moderate precision. Sensitivity analyses reveal architecture-specific failure modes: spline/basis parameters dominate in FastKAN, while base or scaling parameters dominate in EfficientKAN, GRAM, and PyKAN. Vivado HLS estimates on a Xilinx UltraScale+ device further suggest up to 3.32×\times throughput and 7.7×\times lower estimated dynamic energy per inference under W4A4, exposing a residual basis-evaluation tax that motivates basis-aware microarchitecture. QuantKAN is available at https://github.com/OSU-STARLAB/QuantKAN/.
Nov 7, 2025

Quantization Effects of Artificial Neural Networks for Embedded Edge-Computing Applications

This paper examines the use of Quantized Neural Networks (QNNs) for two resource-constrained scientific applications: automated calibration of semiconductor quantum bits (qubits) and scientific particle detectors. We evaluate the trade-offs between Post-Training Quantization (PTQ), Quantization-Aware Training (QAT), and ultra-low-bit Binary Neural Networks (BNNs) with respect to latency and resource usage. Our results demonstrate that PTQ and QAT are easily implementable solutions achieving a four-fold reduction in memory usage for Ushaped CNN (U-Net) architectures, whereas BNNs are better suitable for use cases with challenging latency requirements. For the training of non-differentiable custom BNNs, we propose a novel, hardware-constrained learning approach using Genetic Algorithms (GAs). We showcase a Look-Up Table (LUT)-based BNN architecture suitable for direct conversion to Very High-Speed Integrated Circuit Hardware Description Language (VHDL) via the HCL4BNN framework. This method achieves nanosecond-scale inference latencies at 10 ns to 15 ns without requiring specialized Digital Signal Processor (DSP) or Block RAM (BRAM) resources.