Alternatives and similar repositories for speculative-decoding

Users that are interested in speculative-decoding are comparing it to the libraries listed below

• shreyansh26 / Speculative-Sampling
  Implementation of Speculative Sampling as described in "Accelerating Large Language Model Decoding with Speculative Sampling" by Deepmind
  ☆108Updated last year
• FasterDecoding / REST
  REST: Retrieval-Based Speculative Decoding, NAACL 2024
  ☆212Updated 2 months ago
• FasterDecoding / SnapKV
  ☆290Updated 4 months ago
• hemingkx / Spec-Bench
  Spec-Bench: A Comprehensive Benchmark and Unified Evaluation Platform for Speculative Decoding (ACL 2024 Findings)
  ☆338Updated 7 months ago
• dilab-zju / self-speculative-decoding
  Code associated with the paper **Draft & Verify: Lossless Large Language Model Acceleration via Self-Speculative Decoding**
  ☆210Updated 9 months ago
• yxli2123 / LoftQ
  ☆235Updated last year
• shawntan / scattermoe
  Triton-based implementation of Sparse Mixture of Experts.
  ☆253Updated 2 months ago
• opengear-project / GEAR
  GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM
  ☆170Updated last year
• astramind-ai / Mixture-of-depths
  Unofficial implementation for the paper "Mixture-of-Depths: Dynamically allocating compute in transformer-based language models"
  ☆175Updated last year
• jy-yuan / KIVI
  [ICML 2024] KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache
  ☆335Updated 2 weeks ago
• FMInference / DejaVu
  ☆347Updated last year
• FMInference / H2O
  [NeurIPS'23] H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models.
  ☆487Updated last year
• kssteven418 / BigLittleDecoder
  [NeurIPS'23] Speculative Decoding with Big Little Decoder
  ☆95Updated last year
• jaymody / speculative-sampling
  Simple implementation of Speculative Sampling in NumPy for GPT-2.
  ☆98Updated 2 years ago
• jongwooko / distillm
  Official PyTorch implementation of DistiLLM: Towards Streamlined Distillation for Large Language Models (ICML 2024)
  ☆240Updated 8 months ago
• FasterDecoding / TEAL
  ☆154Updated 9 months ago
• shadowpa0327 / Palu
  [ICLR 2025] Palu: Compressing KV-Cache with Low-Rank Projection
  ☆148Updated 9 months ago
• SqueezeAILab / KVQuant
  [NeurIPS 2024] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization
  ☆391Updated last year
• HArmonizedSS / HASS
  Official Implementation of "Learning Harmonized Representations for Speculative Sampling" (HASS)
  ☆51Updated 8 months ago
• RulinShao / LightSeq
  Official repository for DistFlashAttn: Distributed Memory-efficient Attention for Long-context LLMs Training
  ☆218Updated last year
• zankner / Hydra
  ☆48Updated last year
• henryzhongsc / longctx_bench
  Official implementation for Yuan & Liu & Zhong et al., KV Cache Compression, But What Must We Give in Return? A Comprehensive Benchmark o…
  ☆86Updated 9 months ago
• raymin0223 / fast_robust_early_exit
  Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding (EMNLP 2023 Long)
  ☆64Updated last year
• pan-x-c / EE-LLM
  EE-LLM is a framework for large-scale training and inference of early-exit (EE) large language models (LLMs).
  ☆71Updated last year
• AnswerDotAI / cold-compress
  Cold Compress is a hackable, lightweight, and open-source toolkit for creating and benchmarking cache compression methods built on top of…
  ☆146Updated last year
• mit-han-lab / Quest
  [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference
  ☆356Updated 4 months ago
• thunlp / Ouroboros
  Ouroboros: Speculative Decoding with Large Model Enhanced Drafting (EMNLP 2024 main)
  ☆112Updated 8 months ago
• Equationliu / Kangaroo
  [NeurIPS 2024] The official implementation of "Kangaroo: Lossless Self-Speculative Decoding for Accelerating LLMs via Double Early Exitin…
  ☆63Updated last year
• yuhuixu1993 / qa-lora
  Official PyTorch implementation of QA-LoRA
  ☆145Updated last year
• efeslab / Atom
  [MLSys'24] Atom: Low-bit Quantization for Efficient and Accurate LLM Serving
  ☆331Updated last year