# Kangaroo: Lossless Self-Speculative Decoding via Double Early Exiting

arXiv: https://arxiv.org/abs/2404.18911 · HTML: https://mlsanity.com/paper/2404.18911/

## Key facts

- Cited by 58 papers in the map (data through 14 Oct 2026).
- Placed in Speculative decoding, one of 12 directions in Efficient LLM inference and serving.
- Builds on 30 of its 34 references that are in the map.
- First posted on arXiv 29 Apr 2024.

## Placed in

- Efficient LLM inference and serving / [Speculative decoding](https://mlsanity.com/inference/speculative-decoding/)

## Builds on (most cited first, 30 of 30)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [Mistral 7B](https://mlsanity.com/paper/2310.06825/) (2023)
- [Fast Inference from Transformers via Speculative Decoding](https://mlsanity.com/paper/2211.17192/) (2022)
- [Root Mean Square Layer Normalization](https://mlsanity.com/paper/1910.07467/) (2019)
- [SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models](https://mlsanity.com/paper/2211.10438/) (2022)
- [Accelerating Large Language Model Decoding with Speculative Sampling](https://mlsanity.com/paper/2302.01318/) (2023)
- [Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads](https://mlsanity.com/paper/2401.10774/) (2024)
- [Fast Transformer Decoding: One Write-Head is All You Need](https://mlsanity.com/paper/1911.02150/) (2019)
- [EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty](https://mlsanity.com/paper/2401.15077/) (2024)
- [Blockwise Parallel Decoding for Deep Autoregressive Models](https://mlsanity.com/paper/1811.03115/) (2018)
- [Confident Adaptive Language Modeling](https://mlsanity.com/paper/2207.07061/) (2022)
- [Break the Sequential Dependency of LLM Inference Using Lookahead Decoding](https://mlsanity.com/paper/2402.02057/) (2024)
- [Unlocking Efficiency in Large Language Model Inference: A Comprehensive Survey of Speculative Decoding](https://mlsanity.com/paper/2401.07851/) (2024)
- [A Survey on Efficient Inference for Large Language Models](https://mlsanity.com/paper/2404.14294/) (2024)
- [DistillSpec: Improving Speculative Decoding via Knowledge Distillation](https://mlsanity.com/paper/2310.08461/) (2023)
- [REST: Retrieval-Based Speculative Decoding](https://mlsanity.com/paper/2311.08252/) (2023)
- [SpecTr: Fast Speculative Decoding via Optimal Transport](https://mlsanity.com/paper/2310.15141/) (2023)
- [Hydra: Sequentially-Dependent Draft Heads for Medusa Decoding](https://mlsanity.com/paper/2402.05109/) (2024)
- [Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding](https://mlsanity.com/paper/2310.05424/) (2023)
- [Inference with Reference: Lossless Acceleration of Large Language Models](https://mlsanity.com/paper/2304.04487/) (2023)
- [SPEED: Speculative Pipelined Execution for Efficient Decoding](https://mlsanity.com/paper/2310.12072/) (2023)
- [A Survey on Transformer Compression](https://mlsanity.com/paper/2402.05964/) (2024)
- [You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model](https://mlsanity.com/paper/2211.11152/) (2022)
- [Recurrent Drafter for Fast Speculative Decoding in Large Language Models](https://mlsanity.com/paper/2403.09919/) (2024)
- [Multi-Candidate Speculative Decoding](https://mlsanity.com/paper/2401.06706/) (2024)
- [PanGu-π: Enhancing Language Model Architectures via Nonlinearity Compensation](https://mlsanity.com/paper/2312.17276/) (2023)
- [PanGu-π Pro:Rethinking Optimization and Architecture for Tiny Language Models](https://mlsanity.com/paper/2402.02791/) (2024)

## Built on by (most cited first, 30 of 58)

- [EAGLE-2: Faster Inference of Language Models with Dynamic Draft Trees](https://mlsanity.com/paper/2406.16858/) (2024)
- [EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test](https://mlsanity.com/paper/2503.01840/) (2025)
- [Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRA](https://mlsanity.com/paper/2410.20672/) (2024)
- [SpecDec++: Boosting Speculative Decoding via Adaptive Candidate Lengths](https://mlsanity.com/paper/2405.19715/) (2024)
- [Judge Decoding: Faster Speculative Sampling Requires Going Beyond Model Alignment](https://mlsanity.com/paper/2501.19309/) (2025)
- [Turning Trash into Treasure: Accelerating Inference of Large Language Models with Token Recycling](https://mlsanity.com/paper/2408.08696/) (2024)
- [Merge, Ensemble, and Cooperate! A Survey on Collaborative Strategies in the Era of Large Language Models](https://mlsanity.com/paper/2407.06089/) (2024)
- [Faster Cascades via Speculative Decoding](https://mlsanity.com/paper/2405.19261/) (2024)
- [Large Language Model Inference Acceleration: A Comprehensive Hardware Perspective](https://mlsanity.com/paper/2410.04466/) (2024)
- [Spec-VLA: Speculative Decoding for Vision-Language-Action Models with Relaxed Acceptance](https://mlsanity.com/paper/2507.22424/) (2025)
- [A Survey on Parallel Text Generation: From Parallel Decoding to Diffusion Language Models](https://mlsanity.com/paper/2508.08712/) (2025)
- [PARD: Accelerating LLM Inference with Low-Cost PARallel Draft Model Adaptation](https://mlsanity.com/paper/2504.18583/) (2025)
- [Speculative Decoding and Beyond: An In-Depth Survey of Techniques](https://mlsanity.com/paper/2502.19732/) (2025)
- [ViSpec: Accelerating Vision-Language Models with Vision-Aware Speculative Decoding](https://mlsanity.com/paper/2509.15235/) (2025)
- [AdaEAGLE: Optimizing Speculative Decoding via Explicit Modeling of Adaptive Draft Structures](https://mlsanity.com/paper/2412.18910/) (2024)
- [D-PACE: Dynamic Position-Aware Cross-Entropy for Parallel Speculative Drafting](https://mlsanity.com/paper/2605.18810/) (2026)
- [Interactive Speculative Planning: Enhance Agent Efficiency through Co-design of System and User Interface](https://mlsanity.com/paper/2410.00079/) (2024)
- [DREAM: Drafting with Refined Target Features and Entropy-Adaptive Cross-Attention Fusion for Multimodal Speculative Decoding](https://mlsanity.com/paper/2505.19201/) (2025)
- [FastMTP: Accelerating LLM Inference with Enhanced Multi-Token Prediction](https://mlsanity.com/paper/2509.18362/) (2025)
- [SP-MoE: Speculative Decoding and Prefetching for Accelerating MoE-based Model Inference](https://mlsanity.com/paper/2510.10302/) (2025)
- [CLaSp: In-Context Layer Skip for Self-Speculative Decoding](https://mlsanity.com/paper/2505.24196/) (2025)
- [Self-Speculative Masked Diffusions](https://mlsanity.com/paper/2510.03929/) (2025)
- [A Survey on Parallel Reasoning](https://mlsanity.com/paper/2510.12164/) (2025)
- [The Efficiency vs. Accuracy Trade-off: Optimizing RAG-Enhanced LLM Recommender Systems Using Multi-Head Early Exit](https://mlsanity.com/paper/2501.02173/) (2025)
- [Training-Free Loosely Speculative Decoding: Accepting Semantically Correct Drafts Beyond Exact Match](https://mlsanity.com/paper/2511.22972/) (2025)
- [Gumiho: A Hybrid Architecture to Prioritize Early Tokens in Speculative Decoding](https://mlsanity.com/paper/2503.10135/) (2025)
- [SPIN: Accelerating Large Language Model Inference with Heterogeneous Speculative Models](https://mlsanity.com/paper/2503.15921/) (2025)
- [Draft on the Fly: Adaptive Self-Speculative Decoding using Cosine Similarity](https://mlsanity.com/paper/2410.01028/) (2024)
- [DEL: Context-Aware Dynamic Exit Layer for Efficient Self-Speculative Decoding](https://mlsanity.com/paper/2504.05598/) (2025)
- [AdaDecode: Accelerating LLM Decoding with Adaptive Layer Parallelism](https://mlsanity.com/paper/2506.03700/) (2025)

## Abstract (arXiv)

Speculative decoding has demonstrated its effectiveness in accelerating the inference of large language models while maintaining a consistent sampling distribution. However, the conventional approach of training a separate draft model to achieve a satisfactory token acceptance rate can be costly. Drawing inspiration from early exiting, we propose a novel self-speculative decoding framework Kangaroo, which uses a fixed shallow sub-network as a self-draft model, with the remaining layers serving as the larger target model. We train a lightweight and efficient adapter module on top of the sub-network to bridge the gap between the sub-network and the full model's representation ability. It is noteworthy that the inference latency of the self-draft model may no longer be negligible compared to the large model, necessitating strategies to increase the token acceptance rate while minimizing the drafting steps of the small model. To address this challenge, we introduce an additional early exiting mechanism for generating draft tokens. Specifically, we halt the small model's subsequent prediction during the drafting phase once the confidence level for the current token falls below a certain threshold. Extensive experiments on the Spec-Bench demonstrate the effectiveness of Kangaroo. Under single-sequence verification, Kangaroo achieves speedups up to 1.68× on Spec-Bench, outperforming Medusa-1 with 88.7% fewer additional parameters (67M compared to 591M). The code for Kangaroo is available at https://github.com/Equationliu/Kangaroo.