{"id":20463026,"url":"https://github.com/ryujaehun/alexnet","last_synced_at":"2026-03-11T05:32:05.645Z","repository":{"id":185048941,"uuid":"106690845","full_name":"ryujaehun/alexnet","owner":"ryujaehun","description":"custom implementation alexnet with tensorflow","archived":false,"fork":false,"pushed_at":"2017-10-17T16:18:40.000Z","size":1886,"stargazers_count":20,"open_issues_count":0,"forks_count":8,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-06-07T00:36:48.901Z","etag":null,"topics":["alexnet","cnn","convolutional-neural-networks","imagenet","tensorflow"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ryujaehun.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":"2017-10-12T12:37:43.000Z","updated_at":"2023-06-07T18:30:47.000Z","dependencies_parsed_at":"2023-07-31T15:31:17.863Z","dependency_job_id":null,"html_url":"https://github.com/ryujaehun/alexnet","commit_stats":null,"previous_names":["ryujaehun/alexnet"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/ryujaehun/alexnet","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryujaehun%2Falexnet","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryujaehun%2Falexnet/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryujaehun%2Falexnet/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryujaehun%2Falexnet/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ryujaehun","download_url":"https://codeload.github.com/ryujaehun/alexnet/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryujaehun%2Falexnet/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30372170,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-10T21:41:54.280Z","status":"online","status_checked_at":"2026-03-11T02:00:07.027Z","response_time":84,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"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":["alexnet","cnn","convolutional-neural-networks","imagenet","tensorflow"],"created_at":"2024-11-15T13:08:48.412Z","updated_at":"2026-03-11T05:32:05.619Z","avatar_url":"https://github.com/ryujaehun.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# alexnet\n___\n\n\n\n## about\n\n\u003eAlexNet is the name of a [convolutional neural](https://en.wikipedia.org/wiki/Convolutional_neural_network) network, originally written with [CUDA](https://en.wikipedia.org/wiki/CUDA) to run with [GPU](https://en.wikipedia.org/wiki/GPU) support, which competed in the [ImageNet Large Scale Visual Recognition Challenge](https://en.wikipedia.org/wiki/ImageNet_Large_Scale_Visual_Recognition_Challenge) in 2012. The network achieved a top-5 error of 15.3%, more than 10.8 percentage points ahead of the runner up. AlexNet was designed by the SuperVision group, consisting of Alex Krizhevsky, Geoffrey Hinton, and Ilya Sutskever. -wikipedia\n\n\n## architecture\n\nThe neural network, which has 60 million parameters and 650,000 neurons, consists\nof five convolutional layers, some of which are followed by max-pooling layers,\nand three fully-connected layers with a final 1000-way softmax. To make training\nfaster, we used non-saturating neurons and a very efficient GPU implementation\nof the convolution operation. To reduce overfitting in the fully-connected\nlayers we employed a recently-developed regularization method called “dropout”\nthat proved to be very effective.\n![](https://kratzert.github.io/images/finetune_alexnet/alexnet.png)\n\n## batch normaliztion\n\n[batch normaliztion](https://arxiv.org/abs/1502.03167)is decreasing technical skill,Gradient Vanishing \u0026 Gradient Exploding\n![](http://nmhkahn.github.io/assets/Casestudy-CNN/alex-norm1.png)\n\n\n### k=2,n=5,α=10−4,β=0.75k=2,n=5,α=10−4,β=0.75\n\n ![](https://shuuki4.files.wordpress.com/2016/01/bn1.png)\n ![](https://shuuki4.files.wordpress.com/2016/01/bn2.png)\n\n## optimizer\n\n Apply AdamOptimizer\n ![](http://i.imgur.com/2dKCQHh.gif?1)\n ![](http://i.imgur.com/pD0hWu5.gif?1)\n ![](http://i.imgur.com/NKsFHJb.gif?1)\n\n## requirement\n\n* tensorflow-gpu (ver.1.3.1)\n* cv2 (ver.3.3.0)\n* numpy (ver 1.13.3)\n* scipy (ver 0.19.1)\n\n\n## Usage\n1. Download the image file from the link below.(LSVRC2012 train,val,test,Development kit (Task 1))\n1. untar.(There is a script in `etc`)\n1. Modify  `IMAGENET_PATH` in train.py hyperparameter(maybe you need).\n\n## train\n___\n\n#### From the beginning\n\n```\npython3 train.py\n```\n\n#### resume training\n\n```\npython3 train.py -resume\n```\n\n## test\n\n```\npython3 test.py\n```\n\n## Classify\n\n```\npython classify.py image\n```\n\n## tensorboard\n\n```\ntensorboard --logdir path/to/summary/train/\n```\n\n![](https://galoismilk.org/storage/etc/graph-large_attrs_key=_too_large_attrs\u0026limit_attr_size=1024\u0026run=.png)\n\n\n## TODO\n\n* ~~apply another optimizer ~~\n* ~~apply tensorboard ~~\n* ~~Fit to a GPU~~\n* ~~Application of the technique to the paper~~\n* Eliminate bottlenecks\n\n\n\n## file_architecture\n\n```\nILSVRC 2012 training set folder should be srtuctured like this:\n\t\tILSVRC2012_img_train\n\t\t\t|_n01440764\n\t\t\t|_n01443537\n\t\t\t|_n01484850\n\t\t\t|_n01491361\n\t\t\t|_ ...\n```    \n\n#### you must untar training file `untar.sh`\n\n\n## download\n\n[download LSVRC 2012 image data file](http://www.image-net.org/challenges/LSVRC/2012/nonpub-downloads)\n\n\n\n## Remove log\n\nIf you do not want to see the log at startup\ntrain.py line 97, remove `allow_soft_placement=True, log_device_placement=True`\n\n## references\n\n[optimizer](http://ruder.io/optimizing-gradient-descent/)\n\n[AlexNet training on ImageNet LSVRC 2012](https://github.com/dontfollowmeimcrazy/imagenet)\n\n[Tensorflow Models](https://github.com/tensorflow/models)\n\n[Tensorflow API](https://www.tensorflow.org/versions/r1.2/api_docs/)\n\n## Licence\n\n[MIT Licence](LICENSE)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fryujaehun%2Falexnet","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fryujaehun%2Falexnet","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fryujaehun%2Falexnet/lists"}