{"id":13700476,"url":"https://github.com/dnouri/cuda-convnet","last_synced_at":"2025-07-17T14:36:32.817Z","repository":{"id":8863013,"uuid":"10574406","full_name":"dnouri/cuda-convnet","owner":"dnouri","description":"My fork of Alex Krizhevsky's cuda-convnet from 2013 where I added dropout, among other features.","archived":false,"fork":false,"pushed_at":"2015-01-23T20:38:44.000Z","size":1175,"stargazers_count":260,"open_issues_count":7,"forks_count":147,"subscribers_count":29,"default_branch":"master","last_synced_at":"2025-07-02T17:10:10.065Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"http://code.google.com/p/cuda-convnet/","language":"Cuda","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/dnouri.png","metadata":{"files":{"readme":"README.rst","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2013-06-08T21:19:06.000Z","updated_at":"2025-05-11T10:01:57.000Z","dependencies_parsed_at":"2022-08-24T10:52:24.773Z","dependency_job_id":null,"html_url":"https://github.com/dnouri/cuda-convnet","commit_stats":null,"previous_names":[],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/dnouri/cuda-convnet","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dnouri%2Fcuda-convnet","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dnouri%2Fcuda-convnet/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dnouri%2Fcuda-convnet/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dnouri%2Fcuda-convnet/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/dnouri","download_url":"https://codeload.github.com/dnouri/cuda-convnet/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dnouri%2Fcuda-convnet/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":265617456,"owners_count":23799052,"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":"2024-08-02T20:00:57.145Z","updated_at":"2025-07-17T14:36:32.801Z","avatar_url":"https://github.com/dnouri.png","language":"Cuda","funding_links":[],"categories":["Cuda"],"sub_categories":[],"readme":"This is my fork of the ``cuda-convnet`` convolutional neural network\nimplementation written by Alex Krizhevsky.\n\n``cuda-convnet`` has quite extensive documentation itself.  Find the\n`MAIN DOCUMENTATION HERE \u003chttp://code.google.com/p/cuda-convnet/\u003e`_.\n\n**Update**: A newer version, `cuda-convnet 2\n\u003chttps://code.google.com/p/cuda-convnet2/\u003e`_, has been released by\nAlex.  This fork is still based on the original cuda-convnet.\n\n===================\nAdditional features\n===================\n\nThis document will only describe the small differences between\n``cuda-convnet`` as hosted on Google Code and this version.\n\nDropout\n=======\n\nDropout is a relatively new regularization technique for neural\nnetworks.  See the `Improving neural networks by preventing\nco-adaptation of feature detectors \u003chttp://arxiv.org/abs/1207.0580\u003e`_\nand `Improving Neural Networks with Dropout\n\u003chttp://www.cs.toronto.edu/~nitish/msc_thesis.pdf‎\u003e`_ papers for\ndetails.\n\nTo set a dropout rate for one of our layers, we use the ``dropout``\nparameter in our model's ``layer-params`` configuration file.  For\nexample, we could use dropout for the last layer in the CIFAR example\nby modifying the section for the fc10 layer to look like so::\n\n  [fc10]\n  epsW=0.001\n  epsB=0.002\n  # ...\n  dropout=0.5\n\nIn practice, you'll probably also want to double the number of\n``outputs`` in that layer.\n\n\nCURAND random seeding\n=====================\n\nAn environment variable ``CONVNET_RANDOM_SEED``, if set, will be used\nto set the CURAND library's random seed.  This is important in order\nto get reproducable results.\n\n\nUpdated to work with CUDA via CMake\n===================================\n\nThe build configuration and code has been updated to work with CUDA\nvia CMake. Run ``cmake .`` and then ``make``. If you have an alternative\nBLAS library just set it with for example ``cmake -DBLAS_LIBRARIES=/usr/lib/libcblas.so  .``.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdnouri%2Fcuda-convnet","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fdnouri%2Fcuda-convnet","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdnouri%2Fcuda-convnet/lists"}