TVP: TV is Almost All You Need for Speculative Decoding
cs.LG, cs.CL
Submitted: 2026-10-31
Updated: 2026-10-31
Comments: 25 pages, 2 figures, 14 tables
Code: https://github.com/lightseekorg/TorchSpec
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- PARD-2: Target-Aligned Parallel Draft Model for Dual-Mode Speculative Decoding
- Accelerating Large Language Model Decoding with Speculative Sampling
- DSpark: Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generation
- DBLAST: Dependent Block Drafting for Stochastic Speculative Decoding
- BV Loss: Block Verification-Aware Loss for Block Diffusion Speculative Decoding
- Draft-OPD: On-Policy Distillation for Speculative Draft Models
- When Is a Draft Accepted? A Theory of Acceptance in Speculative Decoding
- D-PACE: Dynamic Position-Aware Cross-Entropy for Parallel Speculative Drafting
- Acceptance-Aware Draft Model Training for Speculative Decoding
- Qwen3 Technical Report
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis