{"id":24241845,"url":"https://github.com/sophiaas/gtc-invariance","last_synced_at":"2025-09-23T09:32:02.027Z","repository":{"id":217218936,"uuid":"743311548","full_name":"sophiaas/gtc-invariance","owner":"sophiaas","description":"Official PyTorch Implementation of \"A General Framework for Robust G-Invariance in G-Equivariant Networks,\" NeurIPS 2023","archived":false,"fork":false,"pushed_at":"2024-01-24T01:51:26.000Z","size":31,"stargazers_count":4,"open_issues_count":0,"forks_count":1,"subscribers_count":2,"default_branch":"main","last_synced_at":"2024-01-30T14:29:12.892Z","etag":null,"topics":["bispectrum","convolutional-neural-networks","deep-learning","equivariance","geometric-deep-learning","group-theory","invariance"],"latest_commit_sha":null,"homepage":"https://arxiv.org/abs/2310.18564","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/sophiaas.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}},"created_at":"2024-01-15T00:03:31.000Z","updated_at":"2024-01-17T21:42:40.000Z","dependencies_parsed_at":"2024-01-15T04:16:07.060Z","dependency_job_id":"3f090549-eef7-4c7c-8eac-4299cbd2832d","html_url":"https://github.com/sophiaas/gtc-invariance","commit_stats":null,"previous_names":["sophiaas/gtc-invariance"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sophiaas%2Fgtc-invariance","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sophiaas%2Fgtc-invariance/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sophiaas%2Fgtc-invariance/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sophiaas%2Fgtc-invariance/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/sophiaas","download_url":"https://codeload.github.com/sophiaas/gtc-invariance/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":233965533,"owners_count":18758361,"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":["bispectrum","convolutional-neural-networks","deep-learning","equivariance","geometric-deep-learning","group-theory","invariance"],"created_at":"2025-01-14T21:50:14.519Z","updated_at":"2025-09-23T09:32:01.545Z","avatar_url":"https://github.com/sophiaas.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# The $G$-Triple Correlation Layer for Robust $G$-Invariance in $G$-Equivariant Networks\n\nThis repository is the official accompaniment to _A General Framework for Robust G-Invariance in G-Equivariant Networks_ (2023) by Sophia Sanborn and Nina Miolane, published in the _Proceedings of the 37th Conference on Neural Information Processing Systems (NeurIPS)._\n\n## Installation\n\nTo install the requirements and package, run:\n\n```\npip install -r requirements.txt\npython install -e .\n```\n\n## Datasets\n\nTo download the datasets:\n\n1. Download the zip file [here](https://drive.google.com/file/d/1zXDnPNlzo5uTfYo97RlKIDHstaWVQD3L/view?usp=sharing).\n2. Place the file in the top node of this directory, i.e. in `gtc-invariance/`.\n3. Run:\n    ```\n    unzip datasets.zip\n    rm -r datasets.zip\n    ```\n\n## Training\n\nThe full set of hyperparameters and training configurations are specified in the config files in the ```configs/``` folder. To train a model on a particular experiment, you will call the following:\n\n```\nscripts/run_data_agent.py --config [name of config]\nscripts/run_train_agent.py --config [name of config]\n```\n\nThe first call will generate the transformed dataset, and the second will train the model on that dataset. The `config` argument should be followed by the name of a particular config file from `configs/experiments`, e.g. `o2mnist_d16_maxpool`. The `.py` extension of the config should be excluded. Each of the configs in the `configs/experiments` folder combines various model, trainer, etc configs also specified in the `configs` folder. The scripts are set up to log the model with [Weights \u0026 Biases](https://wandb.ai/). A user's wandb entity and project directories should be specified in `configs/logger`.\n\n## License\n\nThis repository is licensed under the MIT License.  \n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsophiaas%2Fgtc-invariance","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsophiaas%2Fgtc-invariance","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsophiaas%2Fgtc-invariance/lists"}