ForeSpark: Joint Draft and Verification Budget Scheduling for Parallel Speculative Decoding
cs.DC, cs.AI, cs.LG
Submitted: 2026-11-10
Updated: 2026-11-13
Comments: 21 pages
Code: https://github.com/vllm-project/specula
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Program Synthesis with Large Language Models
- Accelerating Large Language Model Decoding with Speculative Sampling
- Evaluating Large Language Models Trained on Code
- DScale: Scaling Block-Diffusion Speculative Decoding with Adaptive Verification
- FASER: Fine-Grained Phase Management for Speculative Decoding in Dynamic LLM Serving
- DSpark: Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generation
- Training Verifiers to Solve Math Word Problems
- DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
- DeepSeek-V3 Technical Report
- DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
- GLM-5: from Vibe Coding to Agentic Engineering
- DLoop: Looped Speculative Decoding
- JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree Drafting
- Kimi Linear: An Expressive, Efficient Attention Architecture
- AngelSpec: Towards Real-World High Performance Inference with Speculative Decoding
- D-cut: Adaptive Verification Depth Pruning for Batched Speculative Decoding
- TurboSpec: Closed-loop Speculation Control System for Optimizing LLM Serving Goodput
- Fail Fast, Win Big: Rethinking the Drafting Strategy in Speculative Decoding via Diffusion LLMs
- The Synergy of Speculative Decoding and Batching in Serving Large Language Models
- Qwen3 Technical Report
Related papers
- Programmable Probabilistic Computer with 1,000,000 p-bits
- BatchGen: An Architecture for Scalable and Efficient Batch Inference
- Seamless Reconfiguration for DAG-BFT (Technical Report)
- ELDR: Expert-Locality-Aware Decode Routing for PD-Disaggregated MoE Serving
- iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems
- TIRx: A Unified Tile Primitive ML Compiler Across Heterogeneous Backends