{"id":15140738,"url":"https://github.com/deepgraphlearning/diffpack","last_synced_at":"2025-08-22T05:04:32.774Z","repository":{"id":182139023,"uuid":"667984939","full_name":"DeepGraphLearning/DiffPack","owner":"DeepGraphLearning","description":"Implementation of DiffPack: A Torsional Diffusion Model for Autoregressive Protein Side-Chain Packing","archived":false,"fork":false,"pushed_at":"2023-12-04T03:55:16.000Z","size":10923,"stargazers_count":90,"open_issues_count":9,"forks_count":7,"subscribers_count":7,"default_branch":"main","last_synced_at":"2025-08-22T05:03:32.553Z","etag":null,"topics":["deep-learning","diffusion-models","molecule","protein-structure","score-based-generative-modeling","score-matching"],"latest_commit_sha":null,"homepage":"https://arxiv.org/abs/2306.01794","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/DeepGraphLearning.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":"2023-07-18T18:45:55.000Z","updated_at":"2025-08-11T07:13:25.000Z","dependencies_parsed_at":null,"dependency_job_id":"c7645077-aede-4606-8e53-3570aecc425b","html_url":"https://github.com/DeepGraphLearning/DiffPack","commit_stats":{"total_commits":3,"total_committers":2,"mean_commits":1.5,"dds":"0.33333333333333337","last_synced_commit":"81ee51612abfab5af23a8d33fbb081a024f83a1c"},"previous_names":["deepgraphlearning/diffpack"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/DeepGraphLearning/DiffPack","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DeepGraphLearning%2FDiffPack","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DeepGraphLearning%2FDiffPack/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DeepGraphLearning%2FDiffPack/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DeepGraphLearning%2FDiffPack/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/DeepGraphLearning","download_url":"https://codeload.github.com/DeepGraphLearning/DiffPack/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/DeepGraphLearning%2FDiffPack/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":271588743,"owners_count":24785751,"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","status":"online","status_checked_at":"2025-08-22T02:00:08.480Z","response_time":65,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"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":["deep-learning","diffusion-models","molecule","protein-structure","score-based-generative-modeling","score-matching"],"created_at":"2024-09-26T08:40:36.694Z","updated_at":"2025-08-22T05:04:32.746Z","avatar_url":"https://github.com/DeepGraphLearning.png","language":"Python","readme":"# DiffPack: A Torsional Diffusion Model for Autoregressive Protein Side-Chain Packing\n**DiffPack** is a novel torsional diffusion model designed for predicting the conformation of protein side-chains based on their backbones, as introduced in [arxiv link](https://arxiv.org/abs/2306.01794). By learning the joint distribution of side-chain torsional angles through a process of diffusing and denoising on the torsional space, DiffPack significantly improves angle accuracy across various benchmarks for protein side-chain packing. \n\n\n## Installation\nYou can install DiffPack with the following commands, which will install all the dependencies.\n```shell\nconda create -n diffpack python=3.8\nconda activate diffpack\n```\n\n```shell\nconda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia\nconda install pyg -c pyg\nconda install torchdrug -c milagraph -c conda-forge -c pytorch -c pyg\n```\n\n```shell\npip install biopython==1.77\npip install pyyaml\npip install easydict\n```\n![framwork](asset/diffpack.png)\n\n## Model Checkpoints\nWe provide several versions of DiffPack, each with its own configuration and checkpoint:\n\n| Model                                 | Config                                     | Checkpoint            |\n|---------------------------------------|--------------------------------------------|-----------------------|\n| DiffPack (Vanila)                     | [Config](config/inference.yaml)            | [Google Drive Link](https://drive.google.com/file/d/1tZ9ZOjIxq9SxrkdvbLJyLUBbt2P-mksO/view?usp=sharing) |\n | DiffPack (with Confidence Prediction) | [Config](config/inference_confidence.yaml) | [Google Drive Link](https://drive.google.com/file/d/1tZ9ZOjIxq9SxrkdvbLJyLUBbt2P-mksO/view?usp=sharing) |\n\nThe Vanilla version of DiffPack is the base model, \nwhile the version with Confidence Prediction includes an additional feature that estimates the confidence score of the predicted side-chain conformation.\n\nMost of the configuration is specified in the configuration file. We list some important configuration hyperparameters here:\n- `mode`: Backward mode in diffusion process. We use `ode` or `sde` for DiffPack.\n- `annealed_temp`: Annealing temperature in diffusion process. We use `3` for DiffPack. Ideally, higher value corresponds to lower temperature.\n- `num_sample`: Number of samples in diffusion process. Confidence model will decide which sample to use.\n\n## Running DiffPack\nTo use DiffPack for new proteins on your local machine, we provide the necessary configuration files in the `config` folder. \nFor instance, if you have two pdb files `1a3a.pdb` and `1a3b.pdb`, \nyou can run the following command to infer new proteins and save the results in your chosen output folder:\n```shell\npython script/inference.py -c config/inference_confidence.yaml \\\n                           --seed 2023 \\\n                           --output_dir path/to/output \\\n                           --pdb_files 1a3a.pdb 1a3b.pdb ...\n```\nThis command will generate and save the predicted side-chain conformations for the given proteins. \n\n## Retraining DiffPack\nFor those interested in training DiffPack on their own datasets, we will soon release the code and instructions for this process. \nStay tuned for updates!\n\n## Visualization of Results\n![Visualization](asset/result.png)\n\n## License\nThis project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.\n## Citation\nIf you find DiffPack useful in your research or project, please cite our paper:\n```\n@article{zhang2023diffpack,\n  title={DiffPack: A Torsional Diffusion Model for Autoregressive Protein Side-Chain Packing},\n  author={Zhang, Yangtian and Zhang, Zuobai and Zhong, Bozitao and Misra, Sanchit and Tang, Jian},\n  journal={arXiv preprint arXiv:2306.01794},\n  year={2023}\n}\n```\n\n","funding_links":[],"categories":[],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdeepgraphlearning%2Fdiffpack","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fdeepgraphlearning%2Fdiffpack","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdeepgraphlearning%2Fdiffpack/lists"}