dhcode-cpp / NSA-pytorchLinks
DeepSeek Native Sparse Attention pytorch implementation
☆106Updated 2 weeks ago
Alternatives and similar repositories for NSA-pytorch
Users that are interested in NSA-pytorch are comparing it to the libraries listed below
Sorting:
- SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs☆158Updated last month
- qwen-nsa☆79Updated 2 weeks ago
- Efficient Mixture of Experts for LLM Paper List☆140Updated last month
- TransMLA: Multi-Head Latent Attention Is All You Need (NeurIPS 2025 Spotlight)☆389Updated last month
- ☆148Updated 3 months ago
- [ICLR 2025] PEARL: Parallel Speculative Decoding with Adaptive Draft Length☆120Updated 6 months ago
- Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Models☆330Updated 8 months ago
- [ICML 2025] XAttention: Block Sparse Attention with Antidiagonal Scoring☆239Updated 3 months ago
- [ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Training☆242Updated 2 months ago
- siiRL: Shanghai Innovation Institute RL Framework for Advanced LLMs and Multi-Agent Systems☆222Updated this week
- A sparse attention kernel supporting mix sparse patterns☆342Updated 8 months ago
- Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs☆191Updated 3 weeks ago
- Official implementation of "Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding"☆604Updated this week
- ☆148Updated 7 months ago
- ☆69Updated this week
- 青稞Talk☆156Updated this week
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆147Updated 2 weeks ago
- A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attention☆207Updated 2 months ago
- Implementation for FP8/INT8 Rollout for RL training without performence drop.☆261Updated 3 weeks ago
- ☆119Updated 4 months ago
- This repository is the official implementation of "Jakiro: Boosting Speculative Decoding with Decoupled Multi-Head via MoE"☆29Updated 3 weeks ago
- Tiny-DeepSpeed, a minimalistic re-implementation of the DeepSpeed library☆48Updated 2 months ago
- [CoLM'25] The official implementation of the paper <MoA: Mixture of Sparse Attention for Automatic Large Language Model Compression>☆147Updated 3 months ago
- Implementation of FlashAttention in PyTorch☆172Updated 9 months ago
- ☆130Updated 4 months ago
- ☆427Updated 2 months ago
- ☆43Updated last year
- 16-fold memory access reduction with nearly no loss☆105Updated 7 months ago
- Efficient triton implementation of Native Sparse Attention.☆238Updated 5 months ago
- Train speculative decoding models effortlessly and port them smoothly to SGLang serving.☆439Updated this week