Skip to content

@deepseek-ai DeepSeek

Change the repository type filter

All

    Repositories list

    34 repositories

    • DeepEP

      Public
      DeepEP: an efficient expert-parallel communication library
      Cuda
      •
      MIT License
      •1.3k•9.7k•183•74•Updated May 26, 2026May 26, 2026

    • awesome-deepseek-agent

      Public
      282•2.6k•50•74•Updated May 23, 2026May 23, 2026

    • DeepGEMM

      Public
      DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling
      Cuda
      •
      MIT License
      •1k•7.3k•49•24•Updated May 13, 2026May 13, 2026

    • 3FS

      Public
      A high-performance distributed file system designed to address the challenges of AI training and inference workloads.
      distributed-file-system
      distributed-file-system
      C++
      •
      MIT License
      •1.1k•9.9k•118•33•Updated May 7, 2026May 7, 2026

    • FlashMLA

      Public
      FlashMLA: Efficient Multi-head Latent Attention Kernels
      C++
      •
      MIT License
      •1k•13k•66•36•Updated Apr 30, 2026Apr 30, 2026

    • TileKernels

      Public
      A kernel library written in tilelang
      Python
      •
      MIT License
      •131•1.6k•6•7•Updated Apr 23, 2026Apr 23, 2026

    • awesome-deepseek-integration

      Public
      Integrate the DeepSeek API into popular software
      awesomedeepseek
      awesomedeepseek
      Creative Commons Zero v1.0 Universal
      •4.1k•38k•30•48•Updated Feb 23, 2026Feb 23, 2026

    • DeepSeek-OCR-2

      Public
      Visual Causal Flow
      Python
      •
      Apache License 2.0
      •251•2.9k•50•4•Updated Feb 3, 2026Feb 3, 2026

    • DeepSeek-OCR

      Public
      Contexts Optical Compression
      Python
      •
      MIT License
      •2.1k•23k•248•38•Updated Jan 27, 2026Jan 27, 2026

    • DualPipe

      Public
      A bidirectional pipeline parallelism algorithm for computation-communication overlap in DeepSeek V3/R1 training.
      Python
      •
      MIT License
      •325•3k•3•2•Updated Jan 14, 2026Jan 14, 2026

    • Engram

      Public
      Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models
      Python
      •
      Apache License 2.0
      •336•4.4k•14•6•Updated Jan 14, 2026Jan 14, 2026

    • DeepSeek-Math-V2

      Public
      Python
      •
      Apache License 2.0
      •147•1.6k•8•1•Updated Dec 1, 2025Dec 1, 2025

    • LPLB

      Public
      An early research stage expert-parallel load balancer for MoE models based on linear programming.
      Python
      •
      MIT License
      •36•505•1•0•Updated Nov 19, 2025Nov 19, 2025

    • DeepSeek-V3.2-Exp

      Public
      Python
      •
      MIT License
      •173•1.6k•22•7•Updated Nov 18, 2025Nov 18, 2025

    • awesome-deepseek-coder

      Public
      A curated list of open-source projects related to DeepSeek Coder
      213•787•0•0•Updated Nov 11, 2025Nov 11, 2025

    • DeepSeek-Coder

      Public
      DeepSeek Coder: Let the Code Write Itself
      Python
      •
      MIT License
      •2.8k•24k•134•28•Updated Nov 11, 2025Nov 11, 2025

    • DeepSeek-Coder-V2

      Public
      DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence
      MIT License
      •1.1k•6.8k•65•7•Updated Nov 11, 2025Nov 11, 2025

    • DeepSeek-V3

      Public
      Python
      •
      MIT License
      •17k•104k•166•59•Updated Aug 28, 2025Aug 28, 2025

    • DeepSeek-Prover-V2

      Public
      Other
      •100•1.3k•11•2•Updated Jul 18, 2025Jul 18, 2025

    • DeepSeek-R1

      Public
      MIT License
      •12k•92k•19•24•Updated Jun 27, 2025Jun 27, 2025

    • ESFT

      Public
      Expert Specialized Fine-Tuning
      Python
      •
      MIT License
      •263•735•6•1•Updated May 22, 2025May 22, 2025

    • open-infra-index

      Public
      Production-tested AI infrastructure tools for efficient AGI development and community-driven innovation
      Creative Commons Zero v1.0 Universal
      •287•8k•0•1•Updated May 15, 2025May 15, 2025

    • DreamCraft3D

      Public
      [ICLR 2024] Official implementation of DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior
      generative-model3d-creation3d-generation
      generative-model3d-creation3d-generationdiffusion-modelsimage-to-3daigc
      Python
      •
      MIT License
      •358•3k•34•1•Updated Apr 22, 2025Apr 22, 2025

    • EPLB

      Public
      Expert Parallelism Load Balancer
      Python
      •
      MIT License
      •201•1.4k•6•4•Updated Mar 24, 2025Mar 24, 2025

    • profile-data

      Public
      Analyze computation-communication overlap in V3/R1.
      147•1.2k•13•0•Updated Mar 21, 2025Mar 21, 2025

    • smallpond

      Public
      A lightweight data processing framework built on DuckDB and 3FS.
      data-processingduckdb
      data-processingduckdb
      Python
      •
      MIT License
      •442•5k•23•9•Updated Mar 5, 2025Mar 5, 2025

    • DeepSeek-VL2

      Public
      DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding
      Python
      •
      MIT License
      •1.8k•5.3k•101•20•Updated Feb 26, 2025Feb 26, 2025

    • Janus

      Public
      Janus-Series: Unified Multimodal Understanding and Generation Models
      multimodalunified-modelany-to-any
      multimodalunified-modelany-to-anyfoundation-modelsllmvision-language-pretraining
      Python
      •
      MIT License
      •2.2k•18k•159•26•Updated Feb 1, 2025Feb 1, 2025

    • DeepSeek-V2

      Public
      DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
      MIT License
      •541•5k•83•3•Updated Sep 25, 2024Sep 25, 2024

    • DeepSeek-Prover-V1.5

      Public
      Python
      •
      MIT License
      •239•572•8•1•Updated Aug 16, 2024Aug 16, 2024
    ProTip! When viewing an organization's repositories, you can use the props. filter to filter by custom property.