Infini-AI-Lab / Sirius
Sirius, an efficient correction mechanism, which significantly boosts Contextual Sparsity models on reasoning tasks while maintaining its efficiency gain.
☆21Updated 4 months ago
Alternatives and similar repositories for Sirius:
Users that are interested in Sirius are comparing it to the libraries listed below
- 16-fold memory access reduction with nearly no loss☆63Updated 2 months ago
- ☆23Updated 2 months ago
- [NeurIPS 2024] The official implementation of "Kangaroo: Lossless Self-Speculative Decoding for Accelerating LLMs via Double Early Exitin…☆47Updated 6 months ago
- ☆36Updated 4 months ago
- ☆40Updated last month
- [NeurIPS'23] Speculative Decoding with Big Little Decoder☆88Updated 11 months ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆152Updated 6 months ago
- A sparse attention kernel supporting mix sparse patterns☆93Updated 3 months ago
- TidalDecode: A Fast and Accurate LLM Decoding with Position Persistent Sparse Attention☆26Updated last month
- This repo contains the source code for: Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs☆32Updated 5 months ago
- Code for Palu: Compressing KV-Cache with Low-Rank Projection☆63Updated 2 months ago
- [NeurIPS 2024] Fast Best-of-N Decoding via Speculative Rejection☆38Updated 2 months ago
- The official implementation of paper: SimLayerKV: A Simple Framework for Layer-Level KV Cache Reduction.☆42Updated 3 months ago
- ☆35Updated last month
- Pytorch implementation of our paper accepted by ICML 2024 -- CaM: Cache Merging for Memory-efficient LLMs Inference☆29Updated 6 months ago
- Boosting 4-bit inference kernels with 2:4 Sparsity☆64Updated 4 months ago
- Ouroboros: Speculative Decoding with Large Model Enhanced Drafting (EMNLP 2024 main)☆85Updated 3 months ago
- Code for ACL2022 publication Transkimmer: Transformer Learns to Layer-wise Skim☆21Updated 2 years ago
- Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆107Updated last month
- ☆107Updated 3 months ago
- Official implementation for Yuan & Liu & Zhong et al., KV Cache Compression, But What Must We Give in Return? A Comprehensive Benchmark o…☆59Updated 2 weeks ago
- PyTorch implementation of paper "Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline".☆81Updated last year
- An innovative method expediting LLMs via streamlined semi-autoregressive generation and draft verification.☆24Updated 10 months ago
- SQUEEZED ATTENTION: Accelerating Long Prompt LLM Inference☆36Updated last month
- ☆48Updated 8 months ago
- Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding (EMNLP 2023 Long)☆56Updated 3 months ago
- ☆40Updated 10 months ago
- SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs☆21Updated last month
- Official Implementation of SAM-Decoding: Speculative Decoding via Suffix Automaton☆16Updated 3 weeks ago
- Repo for ACL2023 Findings paper "Emergent Modularity in Pre-trained Transformers"☆21Updated last year