HyCo is a hybrid drafting method that assigns depth to a server model’s multi‑token prediction head for execution on a mobile NPU, while alternative branches are handled by a small language model in a single pass. By returning only the hidden state, input embedding, and key‑value entries, HyCo reduces server GPU time per token by 38.6% on Spec‑Bench and 29.2% on 32K‑token summarization, and achieves faster delivery and higher savings than an SLM‑only drafter. The approach also incorporates quantization, a restricted vocabulary, a 1K‑token window, and an ε‑threshold rule to relax exact sampling distribution preservation.
arXiv:2607. 21535v1 Announce Type: new Abstract: Speculative decoding accelerates autoregressive generation by having a cheap draft propose tokens that a target verifies in parallel.
Speculative decoding accelerates autoregressive generation by having a cheap draft propose tokens that a target verifies in parallel. Frontier models increasingly ship a built-in Multi-Token-Prediction (MTP/NEXTN) draft head under the assumption that the draft is negligibly cheap.
arXiv:2610.15417v1 Announce Type: new Abstract: Mixture-of-experts (MoE) models activate only a few experts per token, but when their expert weights are offloaded to host memory, host-to-device (H2D)...
arXiv:2512. 22420v5 Announce Type: replace-cross Abstract: Speculative decoding (SD) accelerates LLM inference by verifying draft tokens in parallel.
arXiv:2607. 24434v1 Announce Type: cross Abstract: Large Mixture-of-Experts (MoE) language models are attractive for end-device deployment because only a small subset of experts is active per token, but their routed expert weights often exceed accelerator memory.
Large Mixture-of-Experts (MoE) language models are attractive for end-device deployment because only a small subset of experts is active per token, but their routed expert weights often exceed accelerator memory. We target latency-critical single-user settings where routed experts are staged on demand from CPU memory to a GPU or from Flash to a mobile NPU.
arXiv:2609.37532v1 Announce Type: cross Abstract: Growing large language model applications demand efficient inference. At high concurrency, block-diffusion speculative decoding suffers from verifica...
The paper introduces X-CoSD, a communication‑efficient framework for collaborative speculative decoding that allows a small on‑device language model to draft tokens while a server‑side large language model verifies them, even when the two models use different vocabularies. X-CoSD employs hybrid resampling to limit distribution exchange to only the common vocabulary, and its enhanced variant X-CoSD‑E further reduces communication by sending only server‑sampled replacement candidates for device verification. Both variants preserve the server model’s distribution and, according to experiments, accelerate token generation without sacrificing quality.
arXiv:2609.24197v1 Announce Type: new Abstract: Speculative decoding losslessly accelerates large language model inference by having a lightweight draft model predict future tokens for verification b...
arXiv:2609.21172v1 Announce Type: new Abstract: Large language models (LLMs) are moving onto mobile devices for increasingly diverse workloads over text, images, video, and audio. These applications...
arXiv:2606. 13740v1 Announce Type: new Abstract: Diffusion large language models (dLLMs) accelerate generation by denoising multiple tokens in parallel, making them attractive for latency-sensitive mobile inference.
The paper introduces AdaptiveSpec, a training‑free speculative decoding method that simultaneously adapts the per‑step verification rule and the draft‑tree shape using signals generated during decoding. It replaces the fixed token‑match rule with a margin‑based threshold and adjusts tree depth, width, and node count based on draft confidence and recent acceptance history, allowing the total draft count to vary. Experiments on SGLang show up to 56% throughput gains over EAGLE‑3 while maintaining 93% of lossless task accuracy on GSM8K, MATH‑500, and HumanEval across three models.