conftrace_

speculative decoding

111 papers

Explore in graph

Also known as

SD

Co-occurring keywords

large language model (13587) inference acceleration (244) inference optimization (219) token generation (47) draft model (21) model compression (3302) model acceleration (80) language model (4599) autoregressive decoding (33) parallel decoding (74)

Papers

FastV-RAG: Towards Fast and Fine-Grained Video QA with Retrieval-Augmented Generation ACL 2026
Calibrated Speculative Decoding: Frequency-Guided Candidate Selection for Efficient Inference ACL 2026
Benchmarking the Energy Savings with Speculative Decoding Strategies EACL 2026
Scaling LLM Speculative Decoding: Non-Autoregressive Forecasting in Large-Batch Scenarios AAAI 2026
Jakiro: Boosting Speculative Decoding via Decoupled MoE ACL 2026
ConfSpec: Efficient Step-Level Speculative Reasoning via Confidence-Gated Verification ACL 2026
Speculative Decoding Speed-of-Light: Optimal Lower Bounds via Branching Random Walks EACL 2026
TABED: Test-Time Adaptive Ensemble Drafting for Robust Speculative Decoding in LVLMs EACL 2026
AdaSpec: Adaptive Multilingual Speculative Decoding with Self-Synthesized Language-Aware Training and Vocabulary Simplification AAAI 2026
Steering Pretrained Drafters During Speculative Decoding AAAI 2026
HCSpec: Two-Tier Horizontal Cascade Speculative Decoding for High-Efficiency Large Language Model Inference ACL 2026
Sparrow: Text-Anchored Window Attention with Visual-Semantic Glimpsing for Speculative Decoding in Video LLMs ACL 2026
HeteroSpec: Leveraging Contextual Heterogeneity for Efficient Speculative Decoding ACL 2026
See the Forest for the Trees: Loosely Speculative Decoding via Visual-Semantic Guidance for Efficient Inference of Video LLMs ACL 2026
KAD: A Framework for Proxy-based Test-time Alignment with Knapsack Approximation Deferral EACL 2026
An Empirical Study of Speculative Decoding for Small Language Models EACL 2026
Look Before You Leap: A Lookahead Reasoning Quality Gate for Speculative Decoding EACL 2026
T-pro 2.0: An Efficient Russian Hybrid-Reasoning Model and Playground EACL 2026
Talon: Breaking the Synchronization Barrier in Speculative Decoding with Hybrid Model-based and Retrieve-based Drafting AAAI 2026
DIAA: A Decoding-Efficient Inference Acceleration Approach for On-Device Large Language Models AAAI 2026
Democratizing LLM Efficiency: From Hyperscale Optimizations to Universal Deployability AAAI 2026
Annealed Relaxation of Speculative Decoding for Faster Autoregressive Image Generation AAAI 2026
LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification ACL 2026
UniSpec: Training-Free Speculative Decoding for Robust LLM Acceleration Across Languages and Hardware ACL 2026
SSSD: Simply-Scalable Speculative Decoding ACL 2026