Quantization
-
WitCert: Sound Runtime Risk Observability and Gating for KV-Cache Quantization
Abstract
KV-cache quantization is validated today by offline benchmark averages; a deployed system cannot tell whether compression is damaging the request it is serving right now. We give it a provably sound runtime meter, a "DTrace for KV quantization": a per-(layer, head, step) upper bound on the total variation between exact and compressed attention. The meter has two tiers: a deterministic band-norm-witness bound, sound for any cache-preserving black-box quantizer and for any query (adaptive-safe, worst-case Cauchy-Schwarz plus RoPE band-unitarity), and a tighter probabilistic certificate for a controlled subtractively-dithered INT8 quantizer under an explicit request-level failure budget (stated for non-adaptive queries; core theorems machine-checked in Lean 4). Three results. Observability: the meter enters SGLang through an environment-guarded patch, and any scheme registered as one tensor function is measured in live serving. Repair: meter-driven gating, risk-ranked where the witness is saturated and certified where it is informative, empirically restores the quality floor at benchmark scale. For example, raw-cast FP8 improves from 22.8 back to 79.7 on hard RULER tasks, with the difference from uncompressed bounded at [+0.0, +0.8] by a paired test. Analysis: aggressive schemes survive on cross-layer error cancellation, not per-step fidelity. In a 28-layer sweep, no single layer's pollution alone loses anything (0/28), and the certified INT8 cache serves 1.88 times more KV tokens at the same memory in SGLang.
Abstract (translated)
URL
https://arxiv.org/abs/2607.28699
PDF
-
Prox: Training-Free FFN Activation Sparsity via Approximate Intermediate-Channel Salience in LLMs
2026-07-30 02:20:01Jinyi Liu, Wei Chen, Pengyu Chen, Xinyi Yuan, Minghe Bai, Guoquan Wu, Jun WeiAbstract
Feed-forward networks (FFNs) dominate memory traffic and computation in large language model (LLM) inference, making them a primary target for activation sparsification. However, existing training-free methods suffer substantial model-quality degradation at high sparsity due to limitations in their channel-selection strategies. We observe that the SwiGLU intermediate state provides a highly effective channel-selection signal, but obtaining it requires costly dense computation. To address this, we present \emph{Prox}, a two-stage training-free framework for sparse SwiGLU FFNs. Prox hinges on the key insight: sparse execution requires only the channel mask induced by the intermediate state, which can be constructed from the magnitude ranking of its entries rather than their exact values. Specifically, Stage 1 uses input sparsity and quantized proxy weights to construct a shared mask; Stage 2 computes the selected channels exactly, enabling sparse execution of all three projections. Across ten LLMs from six model families, Prox outperforms training-free baselines at all sparsity levels, achieves up to a $1.99\times$ end-to-end decoding speedup at 70\% FFN sparsity, and is compatible with quantization and sparse attention.
Abstract (translated)
URL
https://arxiv.org/abs/2607.27591
PDF
-
HiFloat4 Format for End-To-End Reinforcement Learning Post-Training of Large Language Models
2026-07-29 06:28:26Hei Yi Mak, Shadan Golestan, Hoang Le, Mehran Taghian Jazi, Yunke Peng, Yaoyuan Wang, Yao Wang, Junsong Wang, Tianchi Hu, Fengchen He, Guipeng Hu, Tanzila Rahman, Anandharaju Durai RajuAbstract
We present, to our knowledge, the first end-to-end FP4 RL post-training, in which both the rollout and training policies, including their forward and backward passes, operate at 4-bit precision. A systematic study reveals that the dominant source of degradation in FP4 RL is not training-side quantization error but rollout activation quantization: outliers stretch the dynamic range so far that a large number of activation values underflow to zero under FP4. Counterintuitively, restoring the training policy to higher precision while keeping the rollout in FP4 makes accuracy worse than full FP4 baseline, exposing rollout-training mismatch as the principal failure mode and ruling out standard pretraining-style fixes. We address this with Rollout Residual Quantization (Rollout-ResQ): a single residual correction term constrained to a hardware-friendly sparsity pattern, added only to the FP4 rollout matmul -- a lightweight correction that recovers most of the precision lost to outlier-driven underflow without inflating the rollout's compute footprint. On Qwen2.5-3B and Qwen2.5-Math-7B, Rollout-ResQ paired with the HiFloat4 (HiF4) format -- whose three-level hierarchical scaling preserves resolution under FP4's tight 4-bit budget -- closes the accuracy gap to BF16 from 4.9% to 1.1%, bringing fully quantized FP4 RL within striking distance of full precision. Applied to the open-standard MXFP4, the same recipe narrows the gap from 13.6% to 5.3%, revealing that FP4 format choice is a key factor that determines the ceiling on recoverable accuracy. Together, these results establish HiF4 as the enabling format for end-to-end FP4 RL post-training, and Rollout-ResQ as the activation-side mechanism that makes the gap to BF16 closable.
Abstract (translated)
URL
https://arxiv.org/abs/2607.26515
PDF
-
Meshy T2: Fast Native Mesh Generation with Flow Matching
2026-07-28 16:33:54Jiale Xu, Rendong Liang, Yuhao Long, Siyuan Shen, Zangyueyang Xian, Zeyi Xu, Yuanming HuAbstract
Polygonal meshes are the standard surface representation of modern 3D pipelines, and generating high-quality meshes with artist-style topology is essential for film, gaming, and interactive 3D applications. Mainstream approaches serialize a mesh into a token sequence and decode it autoregressively, which is slow at inference and sensitive to error accumulation, making them impractical for interactive asset creation. We present Meshy T2, a fast native mesh generation framework built on flow matching. At its core is a vertex-set mesh VAE that encodes a mesh into one continuous latent token per vertex and decodes vertices, edge connectivity, and face winding order in a single pass, preserving high-precision geometry and artist-authored topology without vertex quantization or welding. Generation proceeds as a coarse-to-fine cascade of two flow-matching models: an image-conditioned voxel flow first sketches the overall shape as a coarse occupancy scaffold, and a mesh flow then populates the scaffold with per-vertex latent tokens, conditioned on the image, the scaffold, and a requested vertex budget. This design delivers three practical capabilities: interactive generation speed through parallel flow-based synthesis; effective face-count control through the requested vertex budget; and native support for multi-part assets, whose components emerge directly from the generated connectivity. In our experiments, Meshy T2 achieves state-of-the-art geometric fidelity and completes end-to-end image-to-mesh generation within a median of 6 seconds, over an order of magnitude faster than autoregressive baselines. Code and weights will be available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2607.28675
PDF
-
Freq-RemoteVAR: Next-Frequency Autoregressive Modeling for Remote Sensing Change Detection
2026-07-28 14:59:56Luqi Gong, Rui Xu, Yue Chen, Chao Li, Jingqi Hong, Xuefeng ZhaoAbstract
Remote sensing change detection aims to identify land-cover changes from bi-temporal images. Most existing methods follow a one-shot dense prediction paradigm, directly regressing a change mask from fused features. However, such approaches overlook the intrinsic frequency characteristics of change patterns. We propose Freq-RemoteVAR, a frequency autoregressive framework that reformulates change detection as a structured generation problem in the frequency domain. Instead of predicting the change mask in a single step, we introduce a next-frequency prediction paradigm, where change information is progressively generated from coarse to fine. We design a frequency-aware mask tokenization strategy that decomposes change supervision into multi-frequency token targets via Fourier transformation and quantization. We develop a Frequency VAR Transformer, which performs causal autoregressive modeling over frequency tokens. The model starts from learned mask queries and progressively predicts frequency-level tokens conditioned on previously generated tokens and bi-temporal image features, effectively capturing long-range dependencies across frequency scales. We introduce Scale-Aligned RoPE Cross Attention (SRCA) module, which aligns frequency-domain mask queries with spatial-domain bi-temporal features under a unified coordinate system, enhancing spatial-frequency consistency during generation. We propose a Change-quality Control module that adaptively modulates the generation process through dynamic normalization, attention biasing, and spatial offset adjustment, thereby suppressing pseudo-change responses and improving robustness. Extensive experiments on CDD, GZ-CD, and LEVIR-CD demonstrate that Freq-RemoteVAR consistently outperforms existing methods, particularly in challenging scenarios with complex appearance variations and noisy disturbances.
Abstract (translated)
URL
https://arxiv.org/abs/2607.25815
PDF
-
How Small Can You Go? A Controlled Study of LoRA Rank, Target Modules, and Quantization Trade-offs for Text-to-SQL on a 60M-Parameter Model
2026-07-28 11:12:03Mahendra Singh Rathor, Anagheem AzzamAbstract
Parameter-efficient fine-tuning (PEFT) and low-bit quantization are now standard tools for adapting language models under tight compute budgets, yet their interaction is most often studied on billion-parameter models where the design space is expensive to explore. We ask a complementary question: on a specific, fully reproducible 60M-parameter encoder-decoder model (T5-small) and a single-table text-to-SQL benchmark (WikiSQL), how much task accuracy does each efficiency knob actually cost? We run a controlled, single-variable study over (i) LoRA rank r in {2, 4, 8, 16, 32}, (ii) the set of adapted modules, and (iii) numerical precision. We report task accuracy alongside system-level metrics including trainable parameters, peak training memory, inference latency, and throughput, and frame adaptation as a constrained trade-off rather than an accuracy-only objective. Our results show that LoRA with r=16 recovers within 11.6 percentage points of full fine-tuning accuracy (59.6% vs. 71.2% exact-match) while training fewer than 1% of parameters and consuming 31% less peak GPU memory. Within this setting, rank beyond r=16 yields no measurable accuracy gain. QLoRA with INT8 and NF4 quantization achieves comparable accuracy (52.8% and 53.2%) at dramatically lower memory cost (0.60 GB each), demonstrating a compelling trade-off for memory-constrained deployments. All code, configurations, and logs are released for full reproducibility.
Abstract (translated)
URL
https://arxiv.org/abs/2607.25583
PDF
-
faster-enhancer.c: A Dependency-Free int8 Runtime for Streaming Speech Enhancement on Commodity CPUs
Abstract
This is an implementation and measurement study of what it costs to run a streaming speech enhancer on a CPU. We port FastEnhancer-Medium at 48 kHz to faster-enhancer.c, a C runtime with six int8 GEMM tiers selected at initialization, leaving architecture and weights untouched. One Apple M2 core reaches 0.069 real-time factor, against 0.230 for the fp32 ONNX Runtime graph on the same machine, a 3.3x speedup. A Galaxy S23+ (Snapdragon 8 Gen 2) reaches 0.096. The speedup comes from specializing every layer of the runtime around one fixed model. Activation ranges are recomputed per frame, so no calibration set is needed; the k=3 convolutions use Winograd F(2,3); cross-stage state is fp16; the GRU and the dequantization epilogues are fused; and nothing is allocated after startup. Over 824 VoiceBank-DEMAND utterances the engine tracks fp32 to within -0.006 PESQ and -0.08 dB SNR. Speed alone does not settle deployment cost. The enhancer holds a fraction of a core for as long as the microphone is open, so its real-time factor is a duty cycle. A benchmark races through a file; an audio callback does not. Pacing to the 6.67 ms deadline costs 4.2x per frame, saves 49% of the energy, and leaves the cheapest core placement missing 96% of its deadlines. All SIMD tiers within an architecture family emit byte-identical output. The runtime is released as a dependency-free library.
Abstract (translated)
URL
https://arxiv.org/abs/2607.25350
PDF
-
VaLiDRec: Variable-Length LLM-Aligned Semantic IDs for Generative Recommendation
2026-07-28 02:29:10Shutong Qiao, Wei Yuan, Tong Chen, Hao Wang, Quoc Viet Hung Nguyen, Hongzhi YinAbstract
Generative recommendation commonly represents items using fixed-length semantic identifiers (SIDs) constructed through clustering and quantization. However, these artificial codes may overcompress item semantics, remain misaligned with pretrained LLM vocabularies, and require costly autoregressive decoding. In light of this, we propose VaLiDRec, a generative recommendation framework based on variable-length, LLM-aligned semantic identifiers. VaLiDRec constructs SIDs directly from informative native LLM vocabulary tokens via token importance estimation, semantic-quality-aware pruning, and collision-aware refinement, allowing identifier lengths to adapt to item semantic complexity. To model user preferences, VaLiDRec incorporates graph-aware soft prompts and reformulates recommendation as token-set prediction with token-level item scoring, eliminating autoregressive SID generation and beam search. Experiments on four real-world datasets show that VaLiDRec consistently outperforms strong sequential and generative recommendation baselines across all evaluation metrics. It further achieves superior zero-shot item cold-start performance and 87.49$\times$ faster inference than LC-Rec. These results demonstrate that LLM-native variable-length semantic identifiers provide a more expressive and efficient paradigm for generative recommendation.
Abstract (translated)
URL
https://arxiv.org/abs/2607.25209
PDF
-
Enhancing Law-Enforcement Audio Transcription: A LoRA-Based Adaptation of Whisper for BWC Footage
2026-07-27 20:03:47Vivek Senthil, Zhiqiang Tao, Ernest Fokou\'eAbstract
Modern policing faces a "visibility paradox" where law enforcement agencies possess petabytes of Body-Worn Camera (BWC) footage that remains largely unutilized for accountability or systemic review due to the prohibitive labor costs of manual transcription. This research presents a framework for adapting the OpenAI Whisper architecture to the unique acoustic and linguistic challenges of the policing environment. By employing Parameter-Efficient Fine-Tuning (PEFT) through Low-Rank Adaptation (LoRA), we address the significant performance degradation observed in zero-shot models when confronted with high-stress scenarios, sirens, and radio interference. Crucially, we demonstrate that this adaptation is feasible on consumer-grade hardware (Acer Nitro local machine with NVIDIA 4GB GTX GPU) using 8-bit quantization and gradient checkpointing. We further integrate these transcriptions into a symbolic reasoning pipeline using a domain-specific ontology to transform raw audio into evidence-linked incident graphs, achieving a 93.7% lexicon mapping rate for the advancement of procedural justice and transparency.
Abstract (translated)
URL
https://arxiv.org/abs/2607.27245
PDF
-
Enabling Fully Integer-Only Inference for Lightweight Detection Transformers
2026-07-27 18:33:18Thanh Cong Le, Michal Szczepanski, Martyna PorebaAbstract
Vision Transformer detectors now approach the accuracy of CNNs but remain difficult to deploy on NPUs and microcontrollers because key components, including deformable attention, feature fusion, and nonlinear activation functions, are not natively compatible with integer arithmetic. Existing quantized detectors either retain operators such as Softmax, GELU, and LayerNorm or focus on heavyweight backbones, leaving lightweight detection transformers without an end-to-end integer implementation. We address this gap with I-LW-DETR, the first fully integer-only lightweight DETR, in which every operation in the forward pass, including transformer nonlinearities, is executed in integer arithmetic. I-LW-DETR is built upon three key components: a scale-preserving split convolution that assigns independent activation scale to each branch of the multi-scale projector; SD-ShiftGELU, a sign-dependent GELU approximation that preserves element-wise behavior while avoiding the accuracy degradation; and a constrained Shiftmax that maintains stable Softmax normalization. Experimental results demonstrate that the proposed quantization pipeline consistently produces efficient fully integer-only models across different model scales. Across all model scales, the proposed pipeline incurs only a moderate accuracy degradation while reducing the model size by approximately $3.6\times$ and the computational cost by more than one order of magnitude.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24981
PDF
-
Stable FP4 Training via Transposition-Invariant Block Quantization
2026-07-27 18:03:34Mehdi Rahimifar, Amin Darabi, Mehran Taghian Jazi, Xing Huang, Yao Wang, Zhijun Tu, Yufei Cui, Yunke Peng, Hongliang LiAbstract
Reducing training precision is a key lever for improving the e ciency of large language model (LLM) training, but pushing beyond FP8 to 4-bit oating point (FP4) remains challenging due to instability during optimization. We identify a fundamental source of this instability in existing microscaling approaches: scale inconsistency induced by tensor transposition. In conventional 1D block quantization, forward and backward passes assign di erent scaling factors to the same values after transposition, leading to biased and unstable gradient updates. To address this issue, we propose a low-precision training framework based on 2D block FP4 quantization, which enforces transposition-invariant scaling and preserves consistency between forward and backward computations. We further combine this with truncation-free scaling and stochastic rounding to control quantization error and maintain unbiased gradients. To handle the sensitivity of attention mechanisms, we adopt MXFP8 quantization for query and key projections, yielding a practical mixed-precision design. We evaluate our method on dense LLMs up to 7B parameters and a 30B Mixture-of-Experts model, trained on up to 100B tokens. Across all settings, our approach achieves stable end-to-end FP4 training and closely matches BF16 performance, with less than 1.3% degradation in perplexity and downstream accuracy. These results demonstrate that enforcing forwardbackward scaling consistency is su cient to enable practical FP4 training at scale, providing a simple and e ective pathway toward more e cient LLM training.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24953
PDF
-
Hierarchical Group-Conditional Conformal Risk Control for Selective Prediction in Language Models
2026-07-27 15:31:59Murilo Salem, Lu\'isa B\"ohm, Daniel Pontes, Anderson FerrugemAbstract
Large language models serve heterogeneous populations structured by domain, topic difficulty, and linguistic style. Conformal risk control (CRC) gives rigorous marginal risk guarantees for selective prediction with abstention, but marginal guarantees do not imply per-group ones: a model can meet the population budget while systematically over-exposing subgroups to errors. Under mild shift in group composition, standard CRC violates the budget in up to 47% of trials. We propose HG-CRC (Hierarchical Group-Conditional CRC), a post-hoc calibration framework enforcing simultaneous risk guarantees across all nodes of a user-defined group hierarchy. It applies a Bonferroni correction over nodes and a leaf-first policy that uses the most specific applicable threshold, falling back to coarser nodes when a finer one is uncertified or rejects the example. It needs only a held-out calibration set, with no retraining. We evaluate on three models (Qwen3-4B, Llama-3.1-8B-Instruct, Gemma-3-4B) and two benchmarks (ARC Challenge, MMLU-Pro) across eight configurations probing IID generalization, heterogeneity, mixture/domain/prompt/difficulty shift, label noise, and quantization. Main result: HG-CRC reaches an empirical 0% violation rate and WGER=0 on ARC Challenge for high-accuracy models (Qwen3-4B, Llama-3.1-8B). At 500 bootstrap trials these zeros are empirical upper bounds (true rate up to 0.6%), not certified. Results are benchmark-specific: on MMLU-Pro these models abstain entirely or (Llama) retain WGER=0.014. Gemma-3-4B, poorly calibrated here, degrades gracefully by abstaining. Participation cost vs. global CRC is 22 to 37 points. Ablations show hierarchical depth clears the budget: removing difficulty level returns violations to about 11%. Bonferroni is needed for the theoretical guarantee, though its empirical effect matters only with many nodes.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24562
PDF
-
Bigger or Cheaper? Scale and Quantization Effects on Uncertainty Signals in Vision-Language Models Under Image Degradation
2026-07-27 13:47:44M M Asif FerdousAbstract
Vision-language models (VLMs) deployed on consumer hardware must decide when to answer and when to defer, and that decision depends on having a confidence signal that tracks correctness. A practitioner with a fixed memory budget faces a choice between a small model at full precision, the same small model quantized, and a larger model quantized into the same footprint -- three configurations that push the confidence signal in opposing directions. We measure, on identical inputs, how model scale and 4-bit quantization affect two confidence signals in the Qwen2-VL family: the confidence a model states in natural language, and its own mean token probability over the answer it generates. Across 5,700 predictions spanning six realistic photographic degradations at three severities, we find that scale sharply improves the model's internal uncertainty signal (mean error-detection AUROC 0.80 to 0.98 from 2B to 7B) while its verbalized confidence stays weak and often at chance (mean 0.61 to 0.69): the gap between what the model knows and what it says widens rather than closes with size. We find that 4-bit quantization is nearly free for accuracy (-1.6 points) but expensive for the confidence signal (internal AUROC 0.95 to 0.80, and the verbalized-confidence parse rate collapses from 99% to 64%). For a fixed memory budget the recommendation is therefore to prefer a larger quantized model over a smaller full-precision one: 7B-4bit gives both the best accuracy and the best uncertainty signal (internal AUROC 0.98) of the three configurations that fit. We frame the results as selective-prediction operating points so they translate directly into a deployment recommendation, and we argue that error-detection AUROC, not calibration error, is the metric that exposes the difference between the two signals.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24440
PDF
-
MXAttention: Data-Free Optimal Scaling and Pre-Normalization Quantization for MXFP4 Attention
2026-07-27 12:53:22Jianlin Yu, Jing Lin, Linghui Kong, Aiyue Chen, Weiyi Sun, Chenyu Zeng, Wangli Lan, Jinxi Li, Zhuo Zheng, Ziyang Yue, Danning Ke, Fei Yi, Tianchi Hu, Yuan Ding, Yiwu Yao, Junsong WangAbstract
The quadratic cost of attention is a major bottleneck in diffusion-based video generation models. MXFP4 attention provides a promising path toward efficient inference, but direct MXFP4 quantization often degrades generation quality due to two numerical issues: the clipping-underflow trade-off from power-of-two scaling and the row-wise normalization error introduced in the softmax loop. We propose MXAttention, a data-free post-training quantization framework for MXFP4 attention. MXAttention introduces two components: Universal Optimal Scaling (UOS), which exploits the periodic structure of power-of-two microscaling to derive a distribution-independent optimal scaling boundary Qmax=7.25 without calibration or search, and Pre-Normalization Quantization (PNQ), which quantizes unnormalized softmax exponentials before row-wise summation to preserve normalization by construction. Experiments on Wan2.2 and HunyuanVideo show that MXAttention closes at least 95% of the VBench Imaging Quality gap between OCP MXFP4 and FP16, substantially improves frame-level similarity, and preserves FP16-level generation quality with less than 0.01 absolute degradation on all reported VBench metrics. MXAttention also achieves performance competitive with strong NVFP4-based baselines with negligible overhead when fused into the attention pipeline. The implementation is publicly available in MindIE-SD.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24377
PDF
-
A Motion-Aware Vector Quantization Framework with Centroid Reuse for Efficient VLA Inference
2026-07-27 08:29:05Zhuoran Song, Haozhe Jiang, Chunyu Qi, Minnan Pei, Gang Li, Xiaoyao Liang, Haibing GuanAbstract
Vision-Language-Action (VLA) models have demonstrated strong potential for embodied AI, yet their high inference latency on GPUs limits real-time deployment. Existing accelerators, such as Dadu-Corki, improve efficiency but treat VLA models as full-precision workloads, leaving substantial redundancy in both memory and computation underexploited. In this paper, we propose VQVLA, an algorithm-hardware co-design framework that accelerates VLA inference by exploiting weight similarity and execution dynamics. We first introduce MotionVQ, a motion-aware vector quantization scheme that dynamically adjusts quantization precision based on the robot's execution state, reducing memory access while preserving task success rate. We then propose a merged-centroid vectorized GEMM paradigm that operates on the codebook-index representation, eliminating redundant multiplications through spatial aggregation and temporal reuse of centroids. To realize these optimizations, we design an accelerator that efficiently supports dynamic precision selection and centroid-reuse computation. Experimental results show that VQVLA achieves 6.5x, 2.8x, 1.9x, 3.3x, and 4.3x speedup over the A100 GPU, Dadu-Corki, LUT-DLA, CodeGEMM, and ShiftAddLLM, respectively, with negligible accuracy degradation.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24148
PDF
-
ACRL: Adaptive Control of Training-Inference Discrepancy for Stable Reinforcement Learning
2026-07-27 07:05:10Wenwu Fan, Qihong Lin, Zhijie Xia, Zhuo Zheng, Sihao Wang, Qiang Chen, Liangsheng ZhuAbstract
Reinforcement Learning (RL) training for Large Language Models (LLMs) often suffers from instability due to the discrepancy between training and inference. This training-inference discrepancy stems from two primary factors: an architectural separation between training and inference engines, and the use of low-precision quantization in inference versus higher-precision computation in training. To address training instability issues caused by high training-inference discrepancy, we present the principles and methods for its adaptive control. We propose Adaptive Control Reinforcement Learning (ACRL), which adaptively maintains the training-inference discrepancy within a reasonable range to ensure stable RL training. Beyond stabilization, ACRL inherently increases policy entropy, thereby enhancing exploration and improving accuracy. The experimental results show that when the inference engine utilizes FP8 quantization, ACRL consistently maintains the training-inference discrepancy within a reasonable range and stabilizes RL training. Furthermore, ACRL not only matches the accuracy of the BF16 baseline but also outperforms importance sampling (IS) fixes.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24062
PDF
-
Research Report on Noise-Shaped One-Bit Coefficients in Discrete Polynomial Fourier Extension
Abstract
This report studies noise-shaped one-bit coefficients in normalized discrete polynomial Fourier extension. For first-order Sigma-Delta quantization, the error is written as $e_k=u_k-q_k=\Delta v_k$ with a uniformly bounded state. Discrete summation by parts then yields variation estimates for complex weights and an $O(N^{-1})$ approximation rate on compact parameter sets. For the parabolic phase $\phi_{x,t}(\xi)=x\xi+t\xi^2$, the bound is expressed through $J(x,t)=\int_0^1 |x+2t\xi|d\xi$, and the uniform $N^{-1}$ rate is shown to be sharp over the admissible input class. Higher-order finite-record identities are derived with all endpoint traces retained. Under endpoint compatibility, or after explicit boundary correction, an $r$th-order noise-shaped error $e=\Delta^r v$ gives $O(N^{-r})$ decay for sufficiently smooth weights and $O(N^{-(r-1+\alpha)})$ decay for $C^{r-1,\alpha}$ weights. Exact $L^2$ orthogonality identities, fourth-moment formulas, local kernel estimates, and oscillatory transfer bounds are also established. Extensions to polynomial phases, multidimensional parameter families, growing observation regions, and correlated state models are included.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24868
PDF
-
Memory Efficient Audio Synthesis with Decoupled Temporal Depth Diffusion Transformers
2026-07-26 19:20:01Dongseong Hwang, Prasanth Yadla, Kaan Elgin, Shifas Padinjaru Veettil, Sivanand Achanta, Dipjyoti Paul, Ramya Rasipuram, Tyler Johnson, Emad Soroush, Chung-Cheng Chiu, Zhifeng ChenAbstract
Siri Expressive Voices synthesize rich, configurable speech in real time and entirely on device, powered by AFM 3 Core Advanced, Apple's most powerful on-device foundation model. This work presents the memory-efficient audio synthesis architecture behind that capability: a detokenizer that converts the semantic audio tokens emitted by the foundation model into high-fidelity audio within the tight compute and memory budget of the Apple Matrix Coprocessor (AMX). We convert semantic audio tokens to a residual vector quantization (RVQ) representation with a three-component design, a streaming encoder, a temporal decoder, and a depth decoder, that systematically decouples temporal and depth processing. A single reusable depth decoder with Diffusion Transformer (DiT)-style stage conditioning generates all RVQ levels autoregressively, replacing the dedicated per-level decoders of prior multi-decoder architectures, while causal sliding window attention with fixed-window key-value caching yields constant memory complexity independent of sequence length. Deployed on the AMX, the detokenizer sustains roughly 10 ms per generation step, about 16x faster than real time, with a peak runtime memory of only 21 MB and 329 MB of on-device assets, enabling continuous streaming synthesis of 20-320 seconds of audio. This constant, small footprint replaces the linear and quadratic memory scaling of conventional transformer- and GAN-based approaches. Ablation studies validate the key architectural components, and audio quality assessment confirms that the architecture maintains synthesis fidelity while achieving efficiency gains over existing methods. Operating at a 1-billion-parameter activation size within AFM 3 Core Advanced, it improves Mean Opinion Score by +0.28 overall (4.15 vs. 3.87) and by +0.42 on conversational speech (4.24 vs. 3.82) over the prior on-device text-to-speech system.
Abstract (translated)
URL
https://arxiv.org/abs/2607.23811
PDF
-
Tokens are All You Need: Dual-purpose Semantic IDs for Achieving LLM-Level I/O Efficiency in recommendation systems
2026-07-26 18:04:19Baolei Li, Yiping Yuan, Yilin Zheng, Likang Yin, Ling Liu, Fabio Soldo, Romer Rosales, Xinyang Yi, Lichan HongAbstract
Large-scale recommendation systems face "Memory Wall" bottlenecks due to massive, dense embedding tables. While generative retrieval uses discrete tokens for IDs, high-dimensional context still relies on inefficient dense formats. Inspired by computer vision data compression, we propose Dual-purpose Semantic IDs to achieve LLM-level I/O efficiency. Our methodology uses hierarchical quantization to condense continuous embeddings into discrete Semantic IDs performing two concurrent roles: (1) Collaborative Identity: modeling user-item interactions via learnable embedding table; and (2) Content Reconstruction: using a lightweight Semantic Decoder for on-the-fly embedding approximation. This approach replaces massive vector storage with on-demand reconstruction, reducing system overhead and data footprints. We demonstrate the efficacy of our framework through offline evaluations and successful online deployment in production-scale ranking and retrieval systems at a major video sharing platform, showing that discrete tokens are indeed all you need for highly efficient, content-rich recommendation.
Abstract (translated)
URL
https://arxiv.org/abs/2607.24865
PDF
-
MixQuant: Adaptive Mixed-Precision Quantization for Large Language Models
2026-07-25 05:10:01Ashitabh Misra, Madhav Agrawal, Arham Jain, Tarek AbdelzaherAbstract
Mixed-precision quantization improves the accuracy of post-training quantization by allocating higher bitwidths to sensitive layers, but existing methods solve the allocation for a single fixed memory budget. In practice the budget varies across deployments and is unknown at calibration time. Adaptive quantization addresses this with one offline calibration that serves any budget, yet current methods score layer sensitivity in a manner that does not consider its dependency on quantization levels of other layers. We show that a layer's sensitivity depends strongly on the bitwidths of its upstream layers and that this dependence shifts the resulting preferred bit allocation. We propose MixQuant, a technique-agnostic adaptive framework that wraps any base quantizer. MixQuant marginalizes each layer's distortion over random quantized upstream configurations to obtain budget-agnostic scores, calibrates the quantizer's parameters on plans the allocator itself produces, and penalizes allocations that leave layers at the lowest bitwidths. A single greedy pass then serves any budget at deployment. Across Llama-3.2-3B, Llama-2-7B, and Mistral-7B under AWQ and GPTQ, MixQuant outperforms adaptive and mixed-precision baselines in every setting, improving average accuracy by up to 8 points and reducing perplexity from 12.43 to 10.70 at the tightest budget, while matching an ILP solver at negligible deployment cost.
Abstract (translated)
URL
https://arxiv.org/abs/2607.23047
PDF
