HyperDFlash: Hyper-Connection-Aligned Block Speculative Decoding with Gated Residual Reduction

• URL: http://arxiv.org/abs/2606.26744v2
• Date: Mon, 29 Jun 2026 08:34:06 GMT
• Title: HyperDFlash: Hyper-Connection-Aligned Block Speculative Decoding with Gated Residual Reduction
• Abstract summary: We present HyperDFlash, a block-parallel speculative decoding framework tailored to DeepSeek-V4's Hyper-Connections (HC)<n>Despite the strong performance of DeepSeek-V4's native Multi-Token Prediction (MTP) module on initial token drafting, its draft accuracy degrades sharply at later positions.<n>Although the original DFlash method supports efficient one-pass block drafting, it cannot be seamlessly adapted to the HC paradigm.
• Score: 8.146414118644069
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We present HyperDFlash, a block-parallel speculative decoding framework tailored to DeepSeek-V4's Hyper-Connections (HC). Despite the strong performance of DeepSeek-V4's native Multi-Token Prediction (MTP) module on initial token drafting, its draft accuracy degrades sharply at later positions, as error accumulation from unverified intermediate tokens harms draft acceptance rates. Although the original DFlash method supports efficient one-pass block drafting, it cannot be seamlessly adapted to the HC paradigm, since DeepSeek-V4's multi-path residual stream induces inherent feature misalignment with conventional drafting designs. To resolve this architectural mismatch, we propose two dedicated, model-aligned optimizations for HC residual streams. First, we adopt pre-collapse residual states as the exclusive conditioning signal, preserving complete multi-path structural information and better aligning the drafter with the target's native prediction pathway. Second, we replace the heavy generic linear compressor with a lightweight gated residual reducer, whose parameters are directly inherited from the target model's built-in hc_head module. This design yields input-aware path aggregation with three orders of magnitude fewer parameters while maintaining precise architectural alignment. We further enhance model training via a targeted KL distillation loss applied to the LM-head, regularizing predictions against the target distribution to improve early draft quality. Extensive experiments across math reasoning, code synthesis, and conversational benchmarks demonstrate that HyperDFlash consistently outperforms both the native MTP baseline and vanilla DFlash adaptation, achieving substantial gains in average accepted draft length and decoding speedup. These results validate HC alignment, gated reduction, and targeted distillation for high-performance speculative decoding.

Related papers

• CURE: Local Uncertainty Repair for Block-Parallel Speculative Decoding [6.268849049716121]
  CURE is a budget-aware dynamic repair tree designed to repair errors at uncertainty focal points.<n>We demonstrate that CURE increases the average accepted length by 4.2-7.5% over parallel baselines without repair.<n>We also provide a plug-and-play repair module compatible with standard parallel drafting frameworks.
  arXiv  Detail & Related papers  (2026-08-01T08:43:34Z)
• DSpark: Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generation [69.8840101036735]
  Speculative decoding accelerates Large Language Model (LLM) inference.<n>Recent parallel drafters efficiently propose long token sequences in a single forward pass, but suffer from rapid acceptance decay due to a lack of inter-token dependencies.<n>We introduce DSpark, a speculative decoding framework that unifies high- throughput parallel generation with adaptive, load-aware verification.
  arXiv  Detail & Related papers  (2026-07-06T14:28:06Z)
• WhiFlash: Accelerating Speculative Decoding with Token-Level Cross-Paradigm Routing [44.656073829954636]
  WhiFlash is a cross-paradigm SD method that unifies autoregressive and diffusion-based parallel drafting under a single token-level controller.<n>We show that WhiFlash achieves significantly higher acceptance lengths, yielding category-specific throughput gains of up to 69.6%.
  arXiv  Detail & Related papers  (2026-06-05T13:23:33Z)
• Accelerating Birkhoff Projection for Manifold-Constrained Hyper-Connections [6.932606401614012]
  We develop an end-to-end acceleration framework for hyper-constrained hyper-connections (mHCs)<n>We focus on the practically important 4x4 Birkhoff projection setting and develop an end-to-end acceleration framework.
  arXiv  Detail & Related papers  (2026-05-26T13:11:08Z)
• $D^3$-RSMDE: 40$\times$ Faster and High-Fidelity Remote Sensing Monocular Depth Estimation [72.9912717963138]
  Real-time, high-fidelity monocular depth estimation from remote sensing imagery is crucial for numerous applications.<n>Although using Vision Transformer (ViT) backbones for dense prediction is fast, they often exhibit poor perceptual quality.<n>We propose Depth Detail Diffusion for Remote Sensing Monocular Depth Estimation ($D3$-RSMDE) to overcome these limitations.<n>$D3$-RSMDE achieves a notable 11.85% reduction in the Learned Perceptual Image Patch Similarity (LPIPS) perceptual metric.
  arXiv  Detail & Related papers  (2026-03-17T10:50:36Z)
• JPmHC Dynamical Isometry via Orthogonal Hyper-Connections [2.4311915994390403]
  JPmHC is a framework that replaces identity skips with a trainable linear mixer acting on n parallel streams.<n>It prevents gradient pathologies and enhances stability.<n>It achieves faster convergence, higher accuracy, and lower computational cost compared to bistochastic baselines.
  arXiv  Detail & Related papers  (2026-02-20T16:06:01Z)
• Coupled Diffusion-Encoder Models for Reconstruction of Flow Fields [11.478292682955669]
  We propose DiffCoder, a coupled framework that integrates a probabilistic diffusion model with a conventional convolutional ResNet encoder.<n>Under aggressive compression, DiffCoder significantly improves the spectral accuracy while VAEs exhibit substantial degradation.<n>Although both methods show comparable relative L2 reconstruction error, DiffCoder better preserves the underlying distributional structure of the flow.
  arXiv  Detail & Related papers  (2026-01-12T19:24:52Z)
• SpecDiff-2: Scaling Diffusion Drafter Alignment For Faster Speculative Decoding [48.96349422252313]
  Speculative decoding has become the standard approach for accelerating Large Language Model (LLM) inference.<n>It exploits a lossless draft-then-verify procedure to circumvent the latency of autoregressive decoding, achieving impressive speed-ups.<n>This paper proposes SpecDiff-2, a novel framework to jointly address these two bottlenecks.
  arXiv  Detail & Related papers  (2025-11-01T16:12:56Z)
• Mirror Speculative Decoding: Breaking the Serial Barrier in LLM Inference [11.957170239588535]
  Speculative decoding accelerates inference by using a draft model to look ahead.<n>Prior methods partially reduce draft cost but either degrade acceptance or introduce overheads that limit scaling.<n>We present Mirror Speculative Decoding (Mirror-SD), an inference algorithm that breaks the latency-acceptance tradeoff.
  arXiv  Detail & Related papers  (2025-10-15T05:22:57Z)
• MPQ-DMv2: Flexible Residual Mixed Precision Quantization for Low-Bit Diffusion Models with Temporal Distillation [74.34220141721231]
  We present MPQ-DMv2, an improved textbfMixed textbfPrecision textbfQuantization framework for extremely low-bit textbfDiffusion textbfModels.
  arXiv  Detail & Related papers  (2025-07-06T08:16:50Z)
• Accelerating Diffusion LLMs via Adaptive Parallel Decoding [60.407727995313074]
  We introduce adaptive parallel decoding (APD), a novel method that dynamically adjusts the number of tokens sampled in parallel.<n>APD provides markedly higher throughput with minimal quality degradations on downstream benchmarks.
  arXiv  Detail & Related papers  (2025-05-31T06:10:10Z)
• Joint Transmit and Pinching Beamforming for Pinching Antenna Systems (PASS): Optimization-Based or Learning-Based? [89.05848771674773]
  A novel antenna system ()-enabled downlink multi-user multiple-input single-output (MISO) framework is proposed.<n>It consists of multiple waveguides, which equip numerous low-cost antennas, named (PAs)<n>The positions of PAs can be reconfigured to both spanning large-scale path and space.
  arXiv  Detail & Related papers  (2025-02-12T18:54:10Z)
• Complexity Matters: Rethinking the Latent Space for Generative Modeling [65.64763873078114]
  In generative modeling, numerous successful approaches leverage a low-dimensional latent space, e.g., Stable Diffusion. In this study, we aim to shed light on this under-explored topic by rethinking the latent space from the perspective of model complexity.
  arXiv  Detail & Related papers  (2023-07-17T07:12:29Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.