{"id":13737583,"url":"https://github.com/sungyubkim/GBML","last_synced_at":"2025-05-08T14:33:05.757Z","repository":{"id":128823829,"uuid":"215322113","full_name":"sungyubkim/GBML","owner":"sungyubkim","description":"A collection of Gradient-Based Meta-Learning Algorithms with pytorch","archived":false,"fork":false,"pushed_at":"2019-12-09T07:20:34.000Z","size":1476,"stargazers_count":61,"open_issues_count":6,"forks_count":8,"subscribers_count":3,"default_branch":"master","last_synced_at":"2024-11-15T06:32:05.906Z","etag":null,"topics":["cavia","few-shot-learning","gradient-based-meta-learning","implicit-maml","maml","meta-learning","meta-learning-algorithms","neumann-approximation","pytorch","reptile"],"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/sungyubkim.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":"2019-10-15T14:38:55.000Z","updated_at":"2024-08-10T02:25:29.000Z","dependencies_parsed_at":"2023-03-30T12:49:26.269Z","dependency_job_id":null,"html_url":"https://github.com/sungyubkim/GBML","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/sungyubkim%2FGBML","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sungyubkim%2FGBML/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sungyubkim%2FGBML/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sungyubkim%2FGBML/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/sungyubkim","download_url":"https://codeload.github.com/sungyubkim/GBML/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253085772,"owners_count":21851697,"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":["cavia","few-shot-learning","gradient-based-meta-learning","implicit-maml","maml","meta-learning","meta-learning-algorithms","neumann-approximation","pytorch","reptile"],"created_at":"2024-08-03T03:01:54.050Z","updated_at":"2025-05-08T14:33:05.374Z","avatar_url":"https://github.com/sungyubkim.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# GBML\nA collection of Gradient-Based Meta-Learning Algorithms with pytorch\n\n* [MAML](http://proceedings.mlr.press/v70/finn17a)\n\n```python\npython3 main.py --alg=MAML\n```\n\n* [Reptile](https://openai.com/blog/reptile/)\n\n```python\npython3 main.py --alg=Reptile\n```\n\n* [CAVIA](http://proceedings.mlr.press/v97/zintgraf19a)\n\n```python\npython3 main.py --alg=CAVIA\n```\n\n\n\n## Results on miniImagenet\n\n* Without pre-trained encoder (Use 64 channels by default. The exceptions are in parentheses)\n\n|                | 5way 1shot          | 5way 1shot (ours) | 5way 5shot          | 5way 5shot (ours) |\n| -------------- | ------------------- | ----------------- | ------------------- | ----------------- |\n| MAML           | 48.70 ± 1.84%       | 49.00 %           | 63.11 ± 0.92%       | 65.18 %           |\n| Reptile        | 47.07 ± 0.26%       | 43.40 %           | 62.74 ± 0.37%       | -                 |\n| CAVIA          | 49.84 ± 0.68% (128) | 50.07 % (64)      | 64.63 ± 0.53% (128) | 64.21 % (64)      |\n| iMAML          | 49.30 ± 1.88%       | -                 | -                   | -                 |\n| Meta-Curvature | 55.73 ± 0.94% (128) | -                 | 70.30 ± 0.72% (128) | -                 |\n\n* With pre-trained encoder (To be implemented.)\n\n|                | 5way 1shot    | 5way 1shot (ours) | 5way 5shot    | 5way 5shot (ours) |\n| -------------- | ------------- | ----------------- | ------------- | ----------------- |\n| Meta-SGD       | 56.58 ± 0.21% | -                 | 68.84 ± 0.19% | -                 |\n| LEO            | 61.76 ± 0.08% | -                 | 77.59 ± 0.12% | -                 |\n| Meta-Curvature | 61.85 ± 0.10% | -                 | 77.02 ± 0.11% | -                 |\n\n## Dependencies\n\n* Python \u003e= 3.6\n* Pytorch \u003e= 1.2\n* [Higher](https://github.com/facebookresearch/higher) \n* [Torchmeta](https://github.com/tristandeleu/pytorch-meta) \n\n\n\n## To do\n\n* Add ~~ResNet~~ and Pre-trained encoder\n* Add iMAML, Meta-Curvature\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsungyubkim%2FGBML","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsungyubkim%2FGBML","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsungyubkim%2FGBML/lists"}