long-context-modeling 开源项目推荐（25 个）

Xnhyacinth /Awesome-LLM-Long-Context-Modeling

必读基于LLM的长上下文建模的论文和博客🔥

Victorwz /LongMem

Official implementation of our NeurIPS 2023 paper "Augmenting Language Models with Long-Term Memory".

FasterDecoding /SnapKV

该项目在 GitHub 上没有填写描述

MozerWang /Loong

[EMNLP 2024 (Oral)] Leave No Document Behind: Benchmarking Long-Context LLMs with Extended Multi-Doc QA

avanturist322 /awesome-memory-vla

🧠 Awesome Memory-VLA: A curated list of Visual-Language-Action models with memory

zhiyuanhubj /LongRecipe

LongRecipe: Recipe for Efficient Long Context Generalization in Large Language Models

DAMO-NLP-SG /CLEX

[ICLR 2024] CLEX: Continuous Length Extrapolation for Large Language Models

YutongWang1216 /DocMTAgent

Code and data releases for the paper -- DelTA: An Online Document-Level Translation Agent Based on Multi-Level Memory

visual-haystacks /vhs_benchmark

🔥 [ICLR 2025] Official Benchmark Toolkits for "Visual Haystacks: A Vision-Centric Needle-In-A-Haystack Benchmark"

MileBench /MileBench

This repo contains evaluation code for the paper "MileBench: Benchmarking MLLMs in Long Context"

ivonajdenkoska /tulip

[ICLR 2025] Official code repository for "TULIP: Token-length Upgraded CLIP"

hrlics /HoPE

[NeurIPS 2025] HoPE: Hybrid of Position Embedding for Long Context Vision-Language Models

visual-haystacks /mirage

🔥 [ICLR 2025] Official PyTorch Model "Visual Haystacks: A Vision-Centric Needle-In-A-Haystack Benchmark"

LCM-Lab /Elastic-Attention

Elastic Attention: Test-time Adaptive Sparsity Ratios for Efficient Transformers

LCM-Lab /context-denoising-training

context denoising training for long-context modeling

megagonlabs /holobench

🫧 Code for Holistic Reasoning with Long-Context LMs: A Benchmark for Database Operations on Massive Textual Data (Maekawa*, Iso* et al.; ICLR 2025)

hrlics /CoPE

CoPE: Clipped RoPE as A Scalable Free Lunch for Long Context LLMs

LCM-Lab /Awesome-Long-Context-Model-Source-Collection

IncludinglLong-context works and tutorials

pszemraj /megalodon-hf

Pure PyTorch + 🤗 Transformers reimplementation of Megalodon (CEMA + chunked attention) - readable, hackable, no CUDA kernels required

zengxyyu /HiCI

HiCI: Hierarchical Construction-Integration for Long-Context Attention