{"id":13784372,"url":"https://github.com/chenlin9/Gaussianization_Flows","last_synced_at":"2025-05-11T19:32:55.103Z","repository":{"id":84357126,"uuid":"251534843","full_name":"chenlin9/Gaussianization_Flows","owner":"chenlin9","description":null,"archived":false,"fork":false,"pushed_at":"2020-05-06T08:43:28.000Z","size":34,"stargazers_count":22,"open_issues_count":1,"forks_count":9,"subscribers_count":2,"default_branch":"master","last_synced_at":"2024-11-17T20:48:30.638Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/chenlin9.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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":"2020-03-31T07:48:51.000Z","updated_at":"2024-11-07T01:13:54.000Z","dependencies_parsed_at":"2023-07-08T12:31:47.301Z","dependency_job_id":null,"html_url":"https://github.com/chenlin9/Gaussianization_Flows","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/chenlin9%2FGaussianization_Flows","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/chenlin9%2FGaussianization_Flows/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/chenlin9%2FGaussianization_Flows/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/chenlin9%2FGaussianization_Flows/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/chenlin9","download_url":"https://codeload.github.com/chenlin9/Gaussianization_Flows/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253621312,"owners_count":21937503,"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":[],"created_at":"2024-08-03T19:00:41.405Z","updated_at":"2025-05-11T19:32:54.720Z","avatar_url":"https://github.com/chenlin9.png","language":"Python","funding_links":[],"categories":["📝 Publications \u003csmall\u003e(60)\u003c/small\u003e"],"sub_categories":[],"readme":"# Gaussianization Flows\nThis repo contains the implementation for [Gaussianization Flows](https://arxiv.org/abs/2003.01941).\n\n-------------------------------------------------------------------------------------\nIterative Gaussianization is a fixed-point iteration procedure that can transform any continuous random vector into a Gaussian one. \nBased on iterative Gaussianization, we propose a new type of normalizing flow model that enables both efficient \ncomputation of likelihoods and efficient inversion for sample generation. We demonstrate that these models, \nnamed Gaussianization flows, are universal approximators for continuous probability distributions under some regularity \nconditions. Because of this guaranteed expressivity, they can capture multimodal target distributions without compromising \nthe efficiency of sample generation. Experimentally, we show that Gaussianization flows achieve better or comparable \nperformance on several tabular datasets compared to other efficiently invertible flow models \nsuch as Real NVP, Glow and FFJORD. In particular, Gaussianization flows are easier to initialize, \ndemonstrate better robustness with respect to different transformations of the training data, \nand generalize better on small training sets.\n\n\n## Dependencies\n\n* PyTorch\n\n* seaborn\n\n## Running Experiments\n### RBIG Experiments\nTo run RBIG experiments, simply run\n``python rbig.py``\n### Tabular Dataset Experiments\nTo download tabular datasets, follow the instructions [here](https://github.com/gpapamak/maf).\n\nTo run the experiments, run\n```\npython tabular_experiment.py --multidim_kernel --usehouseholder\n```\nand specify the dataset and settings by using the flags\n```angular2\n--total_datapoints  --process_size --dataset --layer --epoch --lr --batch_size\n```","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fchenlin9%2FGaussianization_Flows","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fchenlin9%2FGaussianization_Flows","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fchenlin9%2FGaussianization_Flows/lists"}