{"id":22202491,"url":"https://github.com/suzuki-2001/pytorch-provlae","last_synced_at":"2026-01-22T21:24:53.636Z","repository":{"id":262786309,"uuid":"888202808","full_name":"suzuki-2001/pytorch-proVLAE","owner":"suzuki-2001","description":"Pytorch Implementation of proVLAE","archived":false,"fork":false,"pushed_at":"2024-11-21T10:40:02.000Z","size":2133,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":3,"default_branch":"main","last_synced_at":"2024-12-02T16:30:07.019Z","etag":null,"topics":["3dshapes","beta-vae","disentangled-representations","disentanglement","imagenet","mnist","pytorch","vae"],"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/suzuki-2001.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":"2024-11-14T01:43:01.000Z","updated_at":"2024-11-21T10:40:05.000Z","dependencies_parsed_at":"2024-11-21T10:23:48.292Z","dependency_job_id":null,"html_url":"https://github.com/suzuki-2001/pytorch-proVLAE","commit_stats":null,"previous_names":["suzuki-2001/pytorch-provlae"],"tags_count":1,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suzuki-2001%2Fpytorch-proVLAE","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suzuki-2001%2Fpytorch-proVLAE/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suzuki-2001%2Fpytorch-proVLAE/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suzuki-2001%2Fpytorch-proVLAE/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/suzuki-2001","download_url":"https://codeload.github.com/suzuki-2001/pytorch-proVLAE/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":227981853,"owners_count":17850912,"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":["3dshapes","beta-vae","disentangled-representations","disentanglement","imagenet","mnist","pytorch","vae"],"created_at":"2024-12-02T16:28:47.643Z","updated_at":"2026-01-22T21:24:53.605Z","avatar_url":"https://github.com/suzuki-2001.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# pytorch-proVLAE\n\u003cimg src=\"https://img.shields.io/badge/-Python-F9DC3E.svg?logo=python\u0026style=flat\"\u003e [![Format Code](https://github.com/suzuki-2001/pytorch-proVLAE/actions/workflows/black-format.yaml/badge.svg)](https://github.com/suzuki-2001/pytorch-proVLAE/actions/workflows/black-format.yaml)\n[![Validate Mamba Env](https://github.com/suzuki-2001/pytorch-proVLAE/actions/workflows/validate-mamba-env.yaml/badge.svg)](https://github.com/suzuki-2001/pytorch-proVLAE/actions/workflows/validate-mamba-env.yaml)\n\n\u003c/br\u003e\n\nThis is a PyTorch implementation of the paper [PROGRESSIVE LEARNING AND DISENTANGLEMENT OF HIERARCHICAL REPRESENTATIONS](https://openreview.net/forum?id=SJxpsxrYPS) by Zhiyuan et al, [ICLR 2020](https://iclr.cc/virtual_2020/poster_SJxpsxrYPS.html).\nThe official code for proVLAE, implemented in TensorFlow, is available [here](https://github.com/Zhiyuan1991/proVLAE).\n\n\u003c/br\u003e\n\n\u003cimg src=\"./md/shapes3d.gif\" width=\"100%\"\u003e\n\n☝️ **Visualization of results when traversing the latent space (-1.5 to +1.5) of pytorch-proVLAE trained on 3D Shapes.**\n\n\u0026nbsp;\n\n## Installation\nWe recommend using [mamba](https://mamba.readthedocs.io/en/latest/installation/mamba-installation.html) (via [miniforge](https://github.com/conda-forge/miniforge)) for faster installation of dependencies, but you can also use [conda](https://docs.anaconda.com/miniconda/miniconda-install/).\n```bash\ngit clone https://github.com/suzuki-2001/pytorch-proVLAE.git\ncd pytorch-proVLAE\n\nmamba env create -f env.yaml # or conda\nmamba activate torch-provlae\n```\n\n\u0026nbsp;\n\n## Usage\nYou can train pytorch-proVLAE with the following command. Sample hyperparameters and train configuration are provided in [scripts directory](./scripts/).\nIf you have a checkpoint file from a pythorch-proVLAE training, setting the mode argument to \"traverse\" allows you to inspect the latent traversal. Please ensure that the parameter settings match those used for the checkpoint file when running this mode.\n\n\u003c/br\u003e\n\n```bash\n# training with distributed data parallel\n# we tested NVIDIA V100 PCIE 16GB+32GB, NVIDIA A6000 48GB x2\ntorchrun --nproc_per_node=2 --master_port=29501 src/train.py \\\n    --distributed \\\n    --mode seq_train \\\n    --dataset shapes3d \\\n    --optim adamw \\\n    --num_ladders 3 \\\n    --batch_size 128 \\\n    --num_epochs 15 \\\n    --learning_rate 5e-4 \\\n    --beta 8 \\\n    --z_dim 3 \\\n    --coff 0.5 \\\n    --pre_kl \\\n    --hidden_dim 32 \\\n    --fade_in_duration 5000 \\\n    --output_dir ./output/shapes3d/ \\\n    --data_path ./data\n```\n\u0026nbsp;\n\n## License\nThis repository is licensed under the MIT License - see the [LICENSE](./LICENSE) file for details. This follows the licensing of the [original implementation license](https://github.com/Zhiyuan1991/proVLAE/blob/master/LICENSE) by Zhiyuan.\n\n\u0026nbsp;\n\n***\n*This repository is a contribution to [AIST (National Institute of Advanced Industrial Science and Technology)](https://www.aist.go.jp/) project.\n\n[Human Informatics and Interaction Research Institute](https://unit.aist.go.jp/hiiri/), [Neuronrehabilitation Research Group](https://unit.aist.go.jp/hiiri/nrehrg/) \\\nShosuke Suzuki, Ryusuke Hayashi\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsuzuki-2001%2Fpytorch-provlae","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsuzuki-2001%2Fpytorch-provlae","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsuzuki-2001%2Fpytorch-provlae/lists"}