{"id":20749052,"url":"https://github.com/sash-a/codeepneat","last_synced_at":"2025-07-04T13:35:50.657Z","repository":{"id":44563720,"uuid":"190867166","full_name":"sash-a/CoDeepNEAT","owner":"sash-a","description":"An implementation of CoDeepNEAT using pytorch with extensions","archived":false,"fork":false,"pushed_at":"2021-04-28T09:00:06.000Z","size":121797,"stargazers_count":31,"open_issues_count":2,"forks_count":10,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-03-30T09:31:26.857Z","etag":null,"topics":["automl","machine-learning","neat","neat-algorithm","neural-architecture-search","neuroevolution"],"latest_commit_sha":null,"homepage":"","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/sash-a.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}},"created_at":"2019-06-08T09:03:59.000Z","updated_at":"2024-12-01T21:56:08.000Z","dependencies_parsed_at":"2022-09-17T21:51:45.160Z","dependency_job_id":null,"html_url":"https://github.com/sash-a/CoDeepNEAT","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/sash-a%2FCoDeepNEAT","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sash-a%2FCoDeepNEAT/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sash-a%2FCoDeepNEAT/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sash-a%2FCoDeepNEAT/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/sash-a","download_url":"https://codeload.github.com/sash-a/CoDeepNEAT/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":251311332,"owners_count":21569009,"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":["automl","machine-learning","neat","neat-algorithm","neural-architecture-search","neuroevolution"],"created_at":"2024-11-17T08:20:45.590Z","updated_at":"2025-04-28T12:10:39.628Z","avatar_url":"https://github.com/sash-a.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# CoDeepNEAT\n\nAn implementation of implementation of CoDeepNEAT, originally created by Risto Miikkulainen et al. with our own extensions. Implementation details were taken from their [2017](https://arxiv.org/pdf/1703.00548/) and [2019](https://arxiv.org/pdf/1902.06827.pdf) paper.\n\n## Setup\nRequires [conda](https://docs.conda.io/en/latest/) \n```\nconda create -n cdn --file requirements.txt\nconda activate cdn\npip install tarjan wandb  # these are not available from conda\n```  \n\n## Entry points\nDirectory: ```src/main/```  \n```ft.py``` Fully trains a run from evo.py  \n```evo.py``` Does an evolutionary run  \n```batch_run.py``` Running many different configurations all the way from evolution to fully training. (See note below)  \n\n## Config\nAll config options are in ```src/configuration/configuration.py```  \nExample configs are in ```src/configuration/configs``` directory  \n\n## How to run\n```python src/main/evo.py -g 1 -c base```\n\n## Extensions\nExtensions are detailed in the paper linked above\n\n# Paper\n\nIf you use this code, please cite [our paper](https://ieeexplore.ieee.org/abstract/document/9308151):\n\n```\n@INPROCEEDINGS{9308151,\n  author={S. {Acton} and S. {Abramowitz} and L. {Toledo} and G. {Nitschke}},\n  booktitle={2020 IEEE Symposium Series on Computational Intelligence (SSCI)}, \n  title={Efficiently Coevolving Deep Neural Networks and Data Augmentations}, \n  year={2020},\n  volume={},\n  number={},\n  pages={2543-2550},\n  doi={10.1109/SSCI47803.2020.9308151}}\n```\n\n## Results\nFor detailed results see:  \n[convergence](https://app.wandb.ai/codeepneat/cdn_fully_train/reports/CoDeepNEAT-convergence-results--VmlldzoyMTIyMjY?accessToken=86xwfnm0f8tko6spt71oharczveqgv388hzojcuei7g3z4wonshr4uy5n24bbga6)  \n[evolution](https://app.wandb.ai/codeepneat/cdn/reports/CoDeepNEAT-evolution-results--VmlldzoyMTIyNDI?accessToken=bvjugcdbb1qdgn7czmcyxct60cxkgatapk8nldg1bt7gwy4a4kovlavdt5sy7bz5)\n\nThe accuracies obtained on CIFAR-10  \n![results](results.png \"results\")\n\nThe best data augmentations found  \n![best data augmentations](best_da.png \"best data augmentations\")\n\nThe best genotype found. Using config `configuration/configs/experiments/mms_da_pop_25e.json` and a feature multiplier of 5    \n![best genotype](best_geno.png \"best genotype\")\n\nAnd its corresponding phenotype\n![best phenotype](best_pheno.png \"best phenotype\")\n\n\n#### Note about batch runs\nThis system was developed for rapid tuning of CDN's own hyperparameters on a cluster with a limited number of GPUs. It should not be used for normal training as it was created for our very specific case. Rather do a single run on ```evo.py``` and then fully train it with ```ft.py```.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsash-a%2Fcodeepneat","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsash-a%2Fcodeepneat","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsash-a%2Fcodeepneat/lists"}