draft model

An Empirical Study of Speculative Decoding for Small Language Models EACL 2026 Double: Breaking the Acceleration Limit via Double Retrieval Speculative Parallelism ACL 2026 EDSD: Entropy-Driven Design for Faster Speculative Decoding ACL 2026 HCSpec: Two-Tier Horizontal Cascade Speculative Decoding for High-Efficiency Large Language Model Inference ACL 2026 SSSD: Simply-Scalable Speculative Decoding ACL 2026 Accelerated Diffusion Models via Speculative Sampling ICML 2025 Accelerating Autoregressive Speech Synthesis Inference With Speech Speculative Decoding Interspeech 2025 A Drop-In Solution for On-the-Fly Adaptation of Speculative Decoding in Large Language Models ACL 2025 Decoding Speculative Decoding NAACL 2025 Draft Model Knows When to Stop: Self-Verification Speculative Decoding for Long-Form Generation EMNLP 2025 FastDraft: How to Train Your Draft ACL 2025 Faster Speculative Decoding via Effective Draft Decoder with Pruned Candidate Tree ACL 2025 FastVLM: Self-Speculative Decoding for Fast Vision-Language Model Inference IJCNLP 2025 FR-Spec: Accelerating Large-Vocabulary Language Models via Frequency-Ranked Speculative Sampling ACL 2025 Gumiho: A Hybrid Architecture to Prioritize Early Tokens in Speculative Decoding ICML 2025 Hierarchical Speculative Decoding with Dynamic Window NAACL 2025 MagicDec: Breaking the Latency-Throughput Tradeoff for Long Context Generation with Speculative Decoding ICLR 2025 PEARL: Parallel Speculative Decoding with Adaptive Draft Length ICLR 2025 RASD: Retrieval-Augmented Speculative Decoding ACL 2025 Reward-Guided Speculative Decoding for Efficient LLM Reasoning ICML 2025 Speculate, then Collaborate: Fusing Knowledge of Language Models during Decoding ICML 2025 Speculative Streaming: Efficient and Scalable Speculative Decoding with Multi-Stream Attention EMNLP 2025 Cascade Speculative Drafting for Even Faster LLM Inference NeurIPS 2024 Context-Aware Assistant Selection for Improved Inference Acceleration with Large Language Models EMNLP 2024 DistillSpec: Improving Speculative Decoding via Knowledge Distillation ICLR 2024