awesome-repositories.com
Blog
MCP
awesome-repositories.com

Entdecke die besten Open-Source-Repositories mit KI-gestützter Suche.

EntdeckenKuratierte SuchenOpen-Source-AlternativenSelf-hosted SoftwareBlogSitemap
ProjektMCP-ServerÜber unsRanking-MethodikPresse
RechtlichesDatenschutzAGB
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
·

5 Repos

Awesome GitHub RepositoriesCommunication Optimization

Methods for reducing data overhead and synchronization latency between nodes in a distributed training cluster.

Distinct from Distributed Training: Focuses on the communication efficiency and low-precision data transfer between nodes, rather than the general orchestration of training.

Explore 5 awesome GitHub repositories matching artificial intelligence & ml · Communication Optimization. Refine with filters or upvote what's useful.

Awesome Communication Optimization GitHub Repositories

Finde die besten Repos mit KI.Wir suchen mit KI nach den am besten passenden Repositories.
  • microsoft/deepspeedAvatar von microsoft

    microsoft/DeepSpeed

    42,533Auf GitHub ansehen↗

    DeepSpeed is a distributed deep learning optimization library and framework designed for the training and inference of massive AI models. It serves as a model parallelism orchestrator and a toolkit for scaling large language models across multiple GPUs and compute nodes. The project distinguishes itself through 3D parallelism orchestration, which combines data, pipeline, and tensor parallelism. It utilizes ZeRO-based memory partitioning to eliminate redundant storage and employs CPU-offload memory management to move weights and optimizer states to system RAM. Additionally, it provides special

    Optimizes distributed training communication using low-precision techniques to reduce data traffic and overhead.

    Python
    Auf GitHub ansehen↗42,533
  • uber/horovodAvatar von uber

    uber/horovod

    14,686Auf GitHub ansehen↗

    Horovod is a distributed deep learning framework designed to scale machine learning training across multiple GPUs and nodes. It functions as an orchestrator for multi-GPU scaling and a tool for distributed gradient averaging, allowing users to increase compute capacity without rewriting core model logic. The project provides a consistent communication interface that supports multi-framework model distribution across TensorFlow, PyTorch, Keras, and MXNet. It leverages an MPI distributed training library to synchronize gradients across processes using collective communication operations. The s

    Implements communication batching and computation-communication overlapping to reduce overhead and improve scaling efficiency.

    Python
    Auf GitHub ansehen↗14,686
  • horovod/horovodAvatar von horovod

    horovod/horovod

    14,686Auf GitHub ansehen↗

    Horovod is a distributed deep learning framework and gradient synchronizer designed to scale model training across multiple GPUs and compute nodes. It functions as a distributed training orchestrator and an elastic training engine, utilizing an MPI collective communication library to synchronize weights and gradients across TensorFlow, PyTorch, Keras, and MXNet models. The system distinguishes itself through dynamic elastic scaling, which allows it to adjust the number of active workers at runtime and recover from node failures. It optimizes communication efficiency using tensor fusion batchi

    Optimizes collective communication using all-reduce and all-gather operations to synchronize gradients efficiently.

    Python
    Auf GitHub ansehen↗14,686
  • nvidia/ncclAvatar von NVIDIA

    NVIDIA/nccl

    4,816Auf GitHub ansehen↗

    NCCL ist eine Hochleistungs-Kommunikationsbibliothek und ein Framework für verteiltes GPU-Computing, das für die Ausführung kollektiver und Punkt-zu-Punkt-Datenaustausche über mehrere GPUs in Einzel- oder Multi-Node-Systemen entwickelt wurde. Es dient als RDMA-GPU-Transportschicht und Speicher-Orchestrator, der die hochbandbreitige Synchronisation von Daten und Modellgradienten für verteiltes GPU-Training und Inference erleichtert. Die Bibliothek zeichnet sich durch ihre Fähigkeit aus, Kommunikationsprimitive direkt aus GPU-Kernels auszuführen, wodurch die Host-CPU aus dem kritischen Pfad entfernt wird. Sie nutzt topologiebewusste Pfadauswahl zur Optimierung der Datenbewegung und verwendet RDMA-basierten Netzwerktransport, einschließlich InfiniBand und NVLink, um Zero-Copy-Speicherzugriffe zwischen Geräten über verschiedene physische Knoten hinweg zu ermöglichen. Das Projekt deckt eine breite Palette an kollektiven Kommunikationsmustern ab, darunter Reduktionen, Broadcasts, Gathers und All-to-All-Austausche, neben Punkt-zu-Punkt-Remote-Speicherzugriffen. Es bietet umfassendes Communicator-Management für die Initialisierung, Partitionierung und Größenanpassung von GPU-Gruppen sowie spezialisiertes Speichermanagement für das Registrieren von Buffern und das Koordinieren von gemeinsam genutztem Gerätespeicher. Das System enthält eine Suite von Monitoring- und Observability-Tools für Health-Tracking, diagnostisches Logging und Echtzeit-Ereignisüberwachung sowie Integrationsschnittstellen für Machine-Learning-Frameworks, CUDA-Graphs, MPI und Python.

    Allows suspending communicator operations to temporarily release dynamic memory and resume them later.

    C++
    Auf GitHub ansehen↗4,816
  • rlinf/rlinfAvatar von RLinf

    RLinf/RLinf

    2,502Auf GitHub ansehen↗

    RLinf is a distributed reinforcement learning orchestrator and embodied AI training framework. It provides the infrastructure to train vision-language-action models and robotic policies using a combination of reinforcement learning and supervised fine-tuning. The system is designed for scaling workloads across GPU clusters, managing the placement of actors, rollout workers, and environment components. It features a specialized robotics data collection pipeline for gathering teleoperated demonstrations and simulation trajectories into standardized replay buffers, alongside a hardware interface

    Improves efficiency in large-scale MoE training by fusing data transformation and communication during token dispatching.

    Pythonagentic-aiembodied-aireinforcement-learning
    Auf GitHub ansehen↗2,502
  1. Home
  2. Artificial Intelligence & ML
  3. Distributed Training
  4. Communication Optimization

Unter-Tags erkunden

  • Communicator SuspensionsCapabilities for temporarily pausing communication groups to release memory resources. **Distinct from Communication Optimization:** Distinct from general communication optimization by focusing specifically on the suspension and resumption of communicator state for memory management.