{"id":13604285,"url":"https://github.com/MachineLearningSystem/geeps","last_synced_at":"2025-04-11T23:32:12.906Z","repository":{"id":185461787,"uuid":"491117150","full_name":"MachineLearningSystem/geeps","owner":"MachineLearningSystem","description":"GPU-specialized parameter server for GPU machine learning.","archived":false,"fork":true,"pushed_at":"2018-04-05T21:15:12.000Z","size":138,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2024-11-07T08:42:42.883Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":false,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":"cuihenggang/geeps","license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/MachineLearningSystem.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2022-05-11T13:12:27.000Z","updated_at":"2021-10-05T02:51:26.000Z","dependencies_parsed_at":null,"dependency_job_id":"41a619e2-fe00-45f5-a395-8b5b9202e070","html_url":"https://github.com/MachineLearningSystem/geeps","commit_stats":null,"previous_names":["machinelearningsystem/geeps"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MachineLearningSystem%2Fgeeps","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MachineLearningSystem%2Fgeeps/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MachineLearningSystem%2Fgeeps/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MachineLearningSystem%2Fgeeps/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/MachineLearningSystem","download_url":"https://codeload.github.com/MachineLearningSystem/geeps/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248495104,"owners_count":21113570,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-08-01T19:00:42.901Z","updated_at":"2025-04-11T23:32:07.870Z","avatar_url":"https://github.com/MachineLearningSystem.png","language":null,"readme":"# GeePS\n\n[![License](https://img.shields.io/badge/license-BSD-blue.svg)](LICENSE)\n\n[GeePS](https://cuihenggang.github.io/archive/paper/[eurosys16]geeps.pdf) is a parameter server library that scales single-machine GPU machine learning applications (such as Caffe) to a cluster of machines.\n\n\n## Download and build GeePS and Caffe application\n\nRun the following command to download GeePS and (our slightly modified) Caffe:\n\n```\ngit clone --recurse-submodules https://github.com/cuihenggang/geeps.git\n```\n\nIf you use the Ubuntu 14.04 system, you can run the following commands (from geeps root directory) to install the dependencies:\n\n```\n./scripts/install-geeps-deps-ubuntu14.sh\n./scripts/install-caffe-deps-ubuntu14.sh\n```\n\nAlso, please make sure your CUDA library is installed in `/usr/local/cuda`.\n\nAfter installing the dependencies, you can build GeePS by simply running this command from geeps root directory:\n\n```\nscons -j8\n```\n\nYou can then build (our slightly modified) Caffe by first entering the `apps/caffe` directory and then running `make -j8`:\n\n```\ncd apps/caffe\nmake -j8\n```\n\n\n## Caffe's CIFAR-10 example on two machines\n\nYou can run Caffe distributedly across a cluster of machines with GeePS. In this section, we will show you the steps to run Caffe's CIFAR-10 example on two machines.\n\nAll commands in this section are executed from the `apps/caffe` directory:\n\n```\ncd apps/caffe\n```\n\nYou will first need to prepare a machine file as `examples/cifar10/2parts/machinefile`, with each line being the host name of one machine. Since we use two machines in this example, this machine file should have two lines, such as:\n\n```\nhost0\nhost1\n```\n\nWe will use `pdsh` to launch commands on those machines with the `ssh` protocol, so please make sure that you can `ssh` to those machines without password.\n\nWhen you have your machine file in ready, you can run the following command to download and prepare the CIFAR-10 dataset:\n\n```\n./data/cifar10/get_cifar10.sh\n./examples/cifar10/2parts/create_cifar10_pdsh.sh\n```\n\nOur script will partition the datasets into two parts, one for each machine. You can then train an Inception network on it with this command:\n\n```\n./examples/cifar10/2parts/train_inception.sh\n```\n\nPlease look at our [wiki](https://github.com/cuihenggang/geeps/wiki) for more details. Happy training!\n\n\n## Automatic training hyperparameter tuning\n\n[MLtuner-GeePS](https://github.com/cuihenggang/mltuner-geeps) is an extended version of GeePS with automatic training hyperparameter tuning support. It includes a lightweight [MLtuner](https://cuihenggang.github.io/archive/paper/[arxiv]mltuner.pdf) module that automatically tunes the training hyperparameters for distributed ML training (including learning rate, momentum, batch size, data staleness, etc).\n\n\n## Reference Paper\n\nHenggang Cui, Hao Zhang, Gregory R. Ganger, Phillip B. Gibbons, and Eric P. Xing.\n[GeePS: Scalable Deep Learning on Distributed GPUs with a GPU-Specialized Parameter Server](https://cuihenggang.github.io/archive/paper/[eurosys16]geeps.pdf).\nIn ACM European Conference on Computer Systems, 2016 (EuroSys'16).\n\nHenggang Cui, Gregory R. Ganger, and Phillip B. Gibbons.\n[MLtuner: System Support for Automatic Machine Learning Tuning](https://cuihenggang.github.io/archive/paper/[arxiv]mltuner.pdf).\narXiv preprint 1803.07445.\n\n","funding_links":[],"categories":["Paper-Code"],"sub_categories":["Training"],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FMachineLearningSystem%2Fgeeps","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FMachineLearningSystem%2Fgeeps","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FMachineLearningSystem%2Fgeeps/lists"}