Quantization-Aware Distillation

Quantization-aware training (QAT) can recover much of the accuracy lost when large language models are compressed below four bits. Existing re- covery stages, however, are commonly optimized on fixed completions or teacher-generated answers, whereas the deployed quantized model condi- tions on prefixes generated by itself. Quantization errors can therefore move the model into states that are absent from offline recovery data. We introduce OnlineQAT, a two-stage framework that first obtains a usable low-bit initialization through block-wise QAT and then performs on-policy distillation (OPD) on student-generated responses. At each visited pre- fix, a frozen full-precision teacher provides a sampled reverse-KL training signal. On Qwen3-1.7B, OnlineQAT obtains the best average among the compared quantized methods: 57.28 at W3A16 and 32.52 at W2A16, im- proving over ReasoningQAT by 2.90 and 0.44 points, respectively. The results suggest that student-visited states provide a useful recovery signal beyond fixed-completion training, particularly at three bits.

Model compression strategies are widely employed to reduce memory footprint and network complexity, particularly for devices with constrained computational, memory, and energy resources. Prior works that rely on simultaneous conversion from floating-point high-precision (FP32) to integer low-precision (INT4) representations and distillation into smaller models suffer from unstable training and drastic degradation of prediction performance. To address these limitations, we propose a unified framework, known as Progressive Multi-Ancestor Bit-depth Distillation (PMABD), that progressively compresses the network while transferring knowledge through a growing pool of higher-precision ancestor teachers. PMABD generates a sequence of intermediate teachers that each learn from all higher-precision ancestors and jointly supervise the final target student. This multi-ancestor, multi-stage design stabilizes ultra-low-bit quantization by lowering quantization noise profiles across training and ensuring stable quantization. Experiments on CIFAR-10/100 with ResNet-20/32/18, and Tiny-ImageNet with MobileNetV2 show that PMABD outperforms state-of-the-art compression frameworks, results in 1.06$\%$ increase in performance of W2A2 (ResNet-18/CIFAR-100) student model. We show that a saturation-based stopping criterion contributes to improve the performance of our final student.

Large language model (LLM) inference is increasingly moving toward lower precision to realize the throughput of hardware accelerators, but aggressive post-training quantization (PTQ) can degrade model quality. We present QATFactory, an open-source framework for deployment-aligned quantization-aware distillation (QAD) and reinforcement learning (QARL). QATFactory simulates deployment-time quantization while performing matrix multiplications in BF16, allowing models to adapt to quantization noise without requiring training hardware that natively supports the target format; for example, it supports NVFP4 training on H100 GPUs, which lack FP4 Tensor Cores. The framework supports NVFP4, MXFP4, and $\text{llama}.\text{cpp}$'s Q4_K format; dense and mixture-of-experts models; and both full-parameter and LoRA-based training. It exports checkpoints directly to vLLM and $\text{llama}.\text{cpp}$ without an additional lossy conversion step or added inference overhead. With QATFactory, we conduct extensive experiments on models ranging from 8B to 230B parameters and evaluate exported checkpoints in production inference engines. Across models and formats, QAD consistently improves deployed-model quality over strong PTQ baselines. On Qwen3.5-9B, QAD achieves average benchmark accuracies of 68.9% under NVFP4 and 66.0% under MXFP4, outperforming the best PTQ results of 65.4% and 56.4%, respectively. Through our experiments, we found that although both FP4 formats quantize weights and activations at deployment, the best training strategy is format-dependent: NVFP4 generally performs better when only weights are quantized during training, whereas MXFP4 benefits from quantizing both weights and activations. At a fixed training token budget, training on fewer 32K sequences improves average accuracy by 1.9 points over training on more 4K sequences.

Recent generative models have become increasingly powerful, but their inference cost continues to grow. Model quantization offers a promising way to compress these models and accelerate inference. However, at 4 bits, activation quantization is substantially more challenging than weight quantization. Recent post-training quantization (PTQ) and quantization-aware training (QAT) methods have made progress in 4-bit activation quantization by introducing smoothing, SVD branches, rotations, mixed precision, or advanced formats such as NVFP4. These additional operators and data types impose demanding requirements on inference engines and hardware, limiting the broad adoption of low-precision models. Can quantization be achieved using only plain low-bit operators? To answer this question, we propose JustQuant, a simple yet effective framework that moves the complexity of low-bit quantization from deployment-time operators into the training process. We first revisit model quantization from the perspective of knowledge distillation and show that a key reason existing PTQ and QAT methods fail is that they typically exploit supervision at only a single level. We then introduce Theseus QAD, a quantization-aware distillation method that progressively applies multi-level supervision, analogous to the gradual replacement process in the Ship of Theseus. Extensive experiments on DiT and diffusion large language models show two distinct regimes. For smaller models, Theseus QAD can serve as a lightweight warm-up stage that substantially improves subsequent QAT with plain operators, while naive QAD may collapse in the same setting. For larger models, Theseus QAD provides a stronger distillation training path than ordinary QAD. Across both regimes, JustQuant improves low-bit quantization quality while avoiding the complex operators required by many existing PTQ methods.

Quantization-aware distillation (QAD) restores much of the short-form question-answering performance lost to sub-3-bit quantization, yet leaves mathematical and code reasoning substantially impaired. Long generations often degenerate into repetitive loops, exhausting the decoding budget without completing a solution. We trace this gap to quantization-amplified exposure bias: QAD trains on fixed corpus prefixes, while quantization-induced deviations compound along the model's own autoregressive trajectories. To address this mismatch, we introduce an on-policy distillation (OPD) stage that places teacher supervision where the quantized model actually goes. Starting from a QAD checkpoint, the student generates through the quantized forward path used at deployment and receives feedback from a frozen full-precision teacher on its own prefixes, combining dense token-level guidance with task-verifier rewards. Across four models at 2.79 and 1.88 effective bits, OPD raises average BF16 performance retention from 35% to 70% on MATH-500 and from 66% to 91% on HumanEval while preserving short-form performance, with reasoning gains substantially exceeding those of continued teacher-forced QAD in matched-budget comparisons. By coupling QAD's stable low-bit initialization with OPD's on-policy reasoning recovery, our framework provides a comprehensive sub-3-bit solution that preserves broad capabilities while restoring long-form reasoning.

Serving large language models cheaply increasingly means shipping models that are both structurally compressed to a fraction of their parameters and quantized to 4 bits. Together these steps degrade reasoning, mathematics, coding, and long-context behavior enough to require a recovery, or healing, stage before deployment. The default recipe, quantization-aware training (QAT), re-fits the compressed, quantized model to hard labels; in our pipeline it converged slowly and collapsed past its peak. We adopted Quantization-Aware Healing (QAH) instead. Because a structurally compressed model is never independently trained at full precision, its bfloat16 checkpoint is a distillation-recovered approximation of the original; QAH distills the 4-bit student directly from the original, uncompressed model. On a GPT-OSS 120B to 60B to MXFP4 pipeline, the QAH student matches or beats its bfloat16 source on 7 of 9 benchmarks at roughly 4 times less weight memory and half the teacher's parameter count, and is released open-weight as Hypernova-60B. Against a matched QAT baseline it reaches a comparable peak about 7 times faster and stays stable under continued training, without hand-tuned early stopping. We also report deployment lessons, including a large, reproducible quality gap between distributed-training backends. Our aim is a recipe deployable without a multi-week hyper-parameter search.

Quantization-Aware Training (QAT) enables the deployment of quantized models with minimal accuracy degradation. However, in practical scenarios, training labels are often unavailable due to privacy, copyright, or cost constraints. Knowledge Distillation (KD) is a common approach to address this challenge, but we observe that prior work combining QAT with KD suffers from a fundamental limitation: during distillation, the range mismatch between the teacher and the quantized student model induces an unattainable residual, resulting in an irreducible lower bound on the distillation loss. Motivated by this observation, we propose SQuaT (Student-Aware Quantized Teacher Features), a label-free QAT framework with KD that theoretically eliminates this lower bound by applying the student's quantization parameters to quantize the teacher's features during distillation. Through comprehensive experiments across diverse settings, we demonstrate that SQuaT consistently outperforms strong baselines, with particularly pronounced gains in extreme low-bit (e.g., 1- and 2-bit) settings. Furthermore, extensive evaluations across various model design choices show that our approach does not rely on specific architectural assumptions, making it broadly applicable across diverse architectures and quantization settings. The source code is available at https://github.com/lcdbsa522/SQuaT.

Large language models (LLMs) are increasingly constrained by memory capacity, weight bandwidth, and checkpoint storage during deployment. Existing low-bit compression methods mainly follow two directions. Scalar or group-wise quantization is simple and compatible with efficient low-precision kernels, but its representation capacity becomes limited when the target budget approaches 2 bits per weight. Vector-quantized weight compression provides a richer block-level representation, but usually introduces explicit codebooks, index lookup, and additional storage accounting. This paper presents BiSCo-LLM, a codebook-free binary spherical coding framework for extreme low-bit LLM weight compression. The core pipeline is built on three components. First, local weight chunks are mapped onto a unit hypersphere and binarized into compact spherical codes, so that the main payload is a bit-packed sign stream rather than explicit VQ centroids. Second, a residual BSQ stage encodes the reconstruction error left by the base spherical codec, providing an explicit rate-distortion path without stored codebooks. Third, category-wise recovery distillation is performed after replacing each Transformer module category, reducing the mismatch between local weight reconstruction and assembled model behavior. A small 8-bit protected-channel path is used as an auxiliary stabilization mechanism for sensitive channels and is counted separately from the BSQ payload. The reported storage budget includes binary codes, neural decoders, protected-channel payloads, LoRA adapters, and metadata.

Low-bit quantization enables deployment of image restoration (IR) networks on resource-constrained devices, but introduces rounding noise that disproportionately degrades high-frequency regions such as edges and fine textures. Existing knowledge distillation (KD) methods apply distillation signals uniformly across all spatial locations, overlooking the varying reconstruction difficulty across image regions. To address this, we propose SPARK (Spatial Policy-driven Adaptive Reinforcement Learning for Knowledge Distillation), a framework that adaptively allocates distillation effort using a lightweight reinforcement learning (RL) policy network. At each training step, a difficulty feature extractor computes four signals, namely Laplacian variance, pixel variance, student reconstruction error, and teacher-student knowledge gap, which are fed into a compact policy CNN that produces a stochastic spatial weight map to modulate the KD loss during quantization-aware training (QAT). SPARK is IR task-agnostic, adds no inference cost, and integrates into any existing QAT pipeline without architectural changes. Experiments on benchmark datasets demonstrate that SPARK consistently outperforms PTQ, QAT, and state-of-the-art (SOTA) KD approaches across multiple student architectures, achieving reconstruction quality closest to the full-precision teacher under significant computational constraints.

Demand for low-precision inference, including NVFP4-based approaches, has grown as large language models are increasingly deployed in latency and cost constrained production environments. Quantization-aware distillation (QAD) helps recover accuracy lost under low bit quantization by training a quantized student to match the output distribution of a frozen higher precision teacher via a KL-divergence loss. In this work, we first provide a representation level diagnosis of QAD: output matching alone can mask internal degradation, because many intermediate activation geometries can yield similar teacher-aligned logits. Using CKA, we show that KL-only QAD can reduce layerwise representational similarity relative to the BF16 teacher, with especially severe drift in RL-post-trained models. This drift correlates with downstream bottlenecks on reasoning and coding tasks, suggesting that low bit recovery requires preserving internal geometry rather than matching outputs alone. Motivated by this finding, we propose CKA-QAD, a CKA-guided representational alignment method for NVFP4 QAD and low bit LLM accuracy recovery. The method adds a lightweight regularizer that preserves internal representational geometry during distillation by aligning layerwise Gram matrices through CKA. Across Nemotron 3 Nano and Qwen3-4B-Thinking-2507, CKA-QAD substantially improves representational alignment and improves downstream reasoning and coding accuracy with modest training overhead. Our findings position CKA-guided representational alignment as a practical complement to output matching for quantized LLM recovery.

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.

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.

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

Dataset Distillation (DD) compresses large datasets into compact synthetic ones that maintain training performance. However, current methods mainly target sample reduction, with limited consideration of data precision and its impact on efficiency. We propose Quantization-aware Dataset Distillation (QuADD), a unified framework that jointly optimizes dataset compactness and precision under fixed bit budgets. QuADD integrates a differentiable quantization module within the distillation loop, enabling end-to-end co-optimization of synthetic samples and quantization parameters. Guided by the rate-distortion perspective, we empirically analyze how bit allocation between sample count and precision influences learning performance. Our framework supports both uniform and adaptive non-uniform quantization, where the latter learns quantization levels from data to represent information-dense regions better. Experiments on image classification and 3GPP beam management tasks show that QuADD surpasses existing DD and post-quantized baselines in accuracy per bit, establishing a new standard for information-efficient dataset distillation.