{"id":19647065,"url":"https://github.com/idsia/automated-cl","last_synced_at":"2026-03-08T00:32:55.550Z","repository":{"id":210605528,"uuid":"725756024","full_name":"IDSIA/automated-cl","owner":"IDSIA","description":"Official repository for the paper \"Automating Continual Learning\"","archived":false,"fork":false,"pushed_at":"2024-04-01T03:19:23.000Z","size":1096,"stargazers_count":14,"open_issues_count":0,"forks_count":3,"subscribers_count":5,"default_branch":"main","last_synced_at":"2025-04-28T15:41:39.669Z","etag":null,"topics":["continual-learning","cuda","fast-weight-programmers","fast-weights","few-shot-learning","linear-transformers","meta-learning","pytorch","self-referential-learning","self-referential-weight-matrix","transformers"],"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/IDSIA.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":"2023-11-30T20:07:46.000Z","updated_at":"2025-04-23T05:57:49.000Z","dependencies_parsed_at":"2023-12-07T16:39:34.979Z","dependency_job_id":"e9b852f2-ec6c-4eb4-a0c4-114616abd80f","html_url":"https://github.com/IDSIA/automated-cl","commit_stats":{"total_commits":6,"total_committers":1,"mean_commits":6.0,"dds":0.0,"last_synced_commit":"4f112b5b505a9b101d0e1568948539359b472923"},"previous_names":["idsia/automated-cl"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/IDSIA/automated-cl","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/IDSIA%2Fautomated-cl","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/IDSIA%2Fautomated-cl/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/IDSIA%2Fautomated-cl/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/IDSIA%2Fautomated-cl/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/IDSIA","download_url":"https://codeload.github.com/IDSIA/automated-cl/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/IDSIA%2Fautomated-cl/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30238896,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-08T00:30:53.000Z","status":"ssl_error","status_checked_at":"2026-03-08T00:30:44.061Z","response_time":53,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"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":["continual-learning","cuda","fast-weight-programmers","fast-weights","few-shot-learning","linear-transformers","meta-learning","pytorch","self-referential-learning","self-referential-weight-matrix","transformers"],"created_at":"2024-11-11T14:42:14.304Z","updated_at":"2026-03-08T00:32:55.531Z","avatar_url":"https://github.com/IDSIA.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Automated Continual Learning\n\nThis is the official code repository for the paper:\n\n[Automating Continual Learning](https://arxiv.org/abs/2312.00276)\n\n\n\u003cdiv\u003e\n\u003cimg src=\"fig_acl.png\" alt=\"ACL overview\" width=\"80% align=\"middle\"\u003e\n\u003c/div\u003e    \n\nThis codebase is originally forked from [IDSIA/modern-srwm](https://github.com/IDSIA/modern-srwm)\nwhich we modified for continual learning (also including improved practical settings for self-referential weight matrices, e.g., better initialization strategy).\n\nNB: this is research code with many sub-optimal implementations (search for `NB:` in `main.py` for various comments).\n\n\n### Acknowledgement\n\nOur codebase also includes code from other public repositories, e.g.,\n* [tristandeleu/pytorch-meta](https://github.com/tristandeleu/pytorch-meta) for standard few-shot learning data preparation/processing and data-loader implementations.\n(forked and slightly modified code can be found under `torchmeta_local`)\n\n* [khurramjaved96/mrcl](https://github.com/khurramjaved96/mrcl) for the OML baseline (Table 3).\nForked and modified code can be found under `oml_baseline_local`. We downloaded their out-of-the-box Omniglot model from their Google drive from the same repository.\n\n* [GT-RIPL/Continual-Learning-Benchmark](https://github.com/GT-RIPL/Continual-Learning-Benchmark): this is not included here but we modified/used it to produce the results for the 2-task class-incremental setting (Table 3)\n\nas well as other architectural implementations (currently not reported in the paper):\n\n* [lucidrains/mlp-mixer-pytorch](https://github.com/lucidrains/mlp-mixer-pytorch) for MLP mixer.\n\n* [yinboc/few-shot-meta-baseline](https://github.com/yinboc/few-shot-meta-baseline/blob/master/models/resnet12.py) for Res-12.\n\nPlease find LICENSE files/mentions in the corresponding directory/fileheaders.\n\n### Requirements\n\nThe basic requirements are same as the original repository [IDSIA/modern-srwm/supervised_learning](https://github.com/IDSIA/modern-srwm/tree/main/supervised_learning).\nWe used PyTorch `1.10.2+cu102` or `1.11.0` in our experiments but newer versions should also work.\n\n### Training \u0026 Evaluation \n\nExample training and evaluation scripts are provided under `scripts`.\nOur pre-trained model checkpoints can be downloaded from this [Google drive link](https://drive.google.com/file/d/13QWED2TRG-JHF8Hiy4NY461cTrvtCLgc/view?usp=sharing).\n\n\n## BibTex\n```\n@article{irie2023automating,\n  title={Automating Continual Learning},\n  author={Irie, Kazuki and Csord{\\'a}s, R{\\'o}bert and Schmidhuber, J{\\\"u}rgen},\n  journal={Preprint arXiv:2312.00276},\n  year={2023}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fidsia%2Fautomated-cl","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fidsia%2Fautomated-cl","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fidsia%2Fautomated-cl/lists"}