DeepSeek Native Sparse Attention pytorch implementation
β121Dec 17, 2025Updated 11 months ago
Alternatives and similar repositories for NSA-pytorch
Users that are interested in NSA-pytorch are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- qwen-nsaβ87Oct 14, 2025Updated last year
- π³ Efficient Triton implementations for "Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention"β1,030Feb 5, 2026Updated 9 months ago
- Implementation of the sparse attention pattern proposed by the Deepseek team in their "Native Sparse Attention" paperβ817Aug 15, 2025Updated last year
- β48Dec 13, 2025Updated 11 months ago
- PyTorch DeepSeek Sparse Attention (DSA) training & inferenceβ24Oct 14, 2025Updated last year
- Managed Kubernetes at scale on DigitalOcean β’ AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- A lightweight CuTe-based CUDA kernel for DuoAttention, optimized for large language model inference.β12Dec 25, 2025Updated 10 months ago
- β104Feb 11, 2026Updated 9 months ago
- An efficient implementation of the NSA (Native Sparse Attention) kernelβ135Jun 24, 2025Updated last year
- Creating the DeepSeek V3 model from scratchβ30Oct 8, 2026Updated last month
- BigBang-Proton is a LLM pretrained on cross-scale, cross-structure, cross-discipline real-world scientific tasks to construct a scientiβ¦β22Nov 8, 2025Updated last year
- β156Mar 4, 2025Updated last year
- We introduce UltraLLaDA , a scaled variant of LLaDA-8B-Base that extends the context length up to 128K tokens with light-weight post-traiβ¦β15Oct 23, 2025Updated last year
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inferenceβ172Oct 13, 2025Updated last year
- Efficient triton implementation of Native Sparse Attention.β286May 23, 2025Updated last year
- Deploy open-source AI quickly and easily - Special Bonus Offer β’ AdRunpod Hub is built for open source. One-click deployment and autoscaling endpoints without provisioning your own infrastructure.
- High performance RMSNorm Implement by using SM Core Storage(Registers and Shared Memory)β31Jan 22, 2026Updated 9 months ago
- FlashTile is a CUDA Tile IR compiler that is compatible with NVIDIA's tileiras, targeting SM70 through SM121 NVIDIA GPUs.β60Feb 6, 2026Updated 9 months ago
- β255Nov 19, 2025Updated last year
- analyse problems of AI with Math and Codeβ31Jul 28, 2025Updated last year
- A lightweight SGLang implementation built from scratch.β12Jul 29, 2025Updated last year
- Research on DeepSeek Sparse Attentionβ41Oct 8, 2025Updated last year
- [ICLR 2025] DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Headsβ545Feb 10, 2025Updated last year
- β37Aug 7, 2025Updated last year
- [ASP-DAC 2025] LightCL: Compact Continual Learning with Low Memory Footprint For Edge Deviceβ16Apr 2, 2025Updated last year
- AI Agents on DigitalOcean Gradient AI Platform β’ AdBuild production-ready AI agents using customizable tools or access multiple LLMs through a single endpoint. Create custom knowledge bases or connect external data.
- Some funny cute/cuteDSL code snippetsβ35Mar 2, 2026Updated 8 months ago
- [AAAI 2026 Oral] SAMCL: Empowering SAM to Continually Learn from Dynamic Domains with Extreme Storage Efficiencyβ21Mar 25, 2026Updated 7 months ago
- Enable Next-sentence Prediction for Large Language Models with Faster Speed, Higher Accuracy and Longer Contextβ44Aug 16, 2024Updated 2 years ago
- β18Jul 18, 2026Updated 4 months ago
- β46Oct 15, 2025Updated last year
- Visualize the Expert Parallelism Load Balancerβ19Mar 15, 2025Updated last year
- A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attentionβ313Dec 1, 2025Updated 11 months ago
- DeepSeek-V3.2-Exp DSA Warmup Lightning Indexer training operator based on tilelangβ52Nov 19, 2025Updated last year
- MoBA: Mixture of Block Attention for Long-Context LLMsβ2,199Apr 3, 2025Updated last year
- Managed Database hosting by DigitalOcean β’ AdPostgreSQL, MySQL, MongoDB, Kafka, Valkey, and OpenSearch available. Automatically scale up storage and focus on building your apps.
- The official repository for SkyLadder: Better and Faster Pretraining via Context Window Schedulingβ44Dec 29, 2025Updated 10 months ago
- π Efficient implementations for emerging model architecturesβ6,068Updated this week
- β20Nov 14, 2026Updated last week
- β10Nov 18, 2024Updated 2 years ago
- A collection of memory efficient attention operators implemented in the Triton language.β328Updated this week
- Design hardware-friendly model architectures and migrate existing LLMs with minimal performance lossβ520Nov 11, 2026Updated last week
- pytorch implementation of DeepSeek Engramβ19Mar 24, 2026Updated 7 months ago
Are these results useful?