{"id":16360081,"url":"https://github.com/frgfm/torchhawk","last_synced_at":"2026-05-04T14:44:47.190Z","repository":{"id":110056756,"uuid":"191137226","full_name":"frgfm/TorchHawk","owner":"frgfm","description":"A benchmark on popular deep learning tricks on computer vision tasks with famous datasets","archived":false,"fork":false,"pushed_at":"2019-08-10T15:17:03.000Z","size":170,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2024-12-29T17:41:31.370Z","etag":null,"topics":["computer-vision","deep-learning","image-classification","mnist","pytorch"],"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/frgfm.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,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2019-06-10T09:26:05.000Z","updated_at":"2019-08-15T15:22:24.000Z","dependencies_parsed_at":"2023-05-20T21:00:13.555Z","dependency_job_id":null,"html_url":"https://github.com/frgfm/TorchHawk","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/frgfm%2FTorchHawk","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/frgfm%2FTorchHawk/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/frgfm%2FTorchHawk/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/frgfm%2FTorchHawk/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/frgfm","download_url":"https://codeload.github.com/frgfm/TorchHawk/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":239727059,"owners_count":19687096,"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":["computer-vision","deep-learning","image-classification","mnist","pytorch"],"created_at":"2024-10-11T02:10:29.933Z","updated_at":"2025-12-31T19:30:23.658Z","avatar_url":"https://github.com/frgfm.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Torch Hawk\nA benchmark on popular deep learning tricks on computer vision tasks with famous datasets. \n\n\n\n## Installation\n\nThis project was developed in Python 3.7 with PyTorch 1.1. If you have a previous version of PyTorch, please consider switching `torch.utils.tensorboard` dependency to [tensorboardX](https://github.com/lanpa/tensorboardX).\n\n```bash\ngit clone https://github.com/frgfm/TorchHawk.git\ncd TorchHawk\npip install -r requirements.txt\nmkdir img_classification/logs\n```\n\n\n\n## Usage\n\nHow to train your model\n\n\n\n### Start your training\n\n\n\n```bash\npython img_classification/main.py lenet5 --dataset mnist -n 10 --lr 5e-4 --batch_size 8 --optimizer adam --drop_rate 0.3 --workers 8\n```\n\n\n\nDepending on your seed, you should get similar results to:\n\n```bash\nData transformations:\nCompose(\n    RandomRotation(degrees=(-30, 30), resample=False, expand=False)\n    ToTensor()\n    Normalize(mean=(0.1307,), std=(0.3081,))\n)\n\nEpoch 1/10 - Validation loss: 0.08775 (Acc@1: 96.97%)                                                                                                           \nValidation loss decreased inf --\u003e 0.08775: saving state...\nEpoch 2/10 - Validation loss: 0.05491 (Acc@1: 98.20%)                                                                                                           \nValidation loss decreased 0.08775 --\u003e 0.05491: saving state...\nEpoch 3/10 - Validation loss: 0.06731 (Acc@1: 97.83%)                                                                                                           \nEpoch     2: reducing learning rate of group 0 to 2.5000e-04.\nEpoch 4/10 - Validation loss: 0.04157 (Acc@1: 98.69%)                                                                                                           \nValidation loss decreased 0.05491 --\u003e 0.04157: saving state...\nEpoch 5/10 - Validation loss: 0.03381 (Acc@1: 98.85%)                                                                                                           \nValidation loss decreased 0.04157 --\u003e 0.03381: saving state...\nEpoch 6/10 - Validation loss: 0.03288 (Acc@1: 98.97%)                                                                                                           \nValidation loss decreased 0.03381 --\u003e 0.03288: saving state...\nEpoch 7/10 - Validation loss: 0.03146 (Acc@1: 99.07%)                                                                                                           \nValidation loss decreased 0.03288 --\u003e 0.03146: saving state...\nEpoch 8/10 - Validation loss: 0.03196 (Acc@1: 99.03%)                                                                                                           \nEpoch     7: reducing learning rate of group 0 to 1.2500e-04.\nEpoch 9/10 - Validation loss: 0.02716 (Acc@1: 99.26%)                                                                                                           \nValidation loss decreased 0.03146 --\u003e 0.02716: saving state...\nEpoch 10/10 - Validation loss: 0.02684 (Acc@1: 99.22%)                                                                                                          \nValidation loss decreased 0.02716 --\u003e 0.02684: saving state...\n\n```\n\n\n\n### Running the tensorboard interface\n\nStart the tensorboard server locally to visualize your training losses:\n\n```bash\ntensorboard --logdir=training-outputs/logs\n```\n\nThen open a new tab in your browser and navigate to `\u003cYOUR_COMPUTER_NAME\u003e:6006`  to monitor your training.\n\n![tb_loss](static/images/tb_loss.png)\n\n\n\n\n\n## Submitting a request / Reporting an issue\n\nRegarding issues, use the following format for the title:\n\n\u003e [Topic] Your Issue name\n\nExample:\n\n\u003e [State saving] Add a feature to automatically save and load model states","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffrgfm%2Ftorchhawk","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ffrgfm%2Ftorchhawk","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffrgfm%2Ftorchhawk/lists"}