{"id":30848312,"url":"https://github.com/arthur-kitsuragi/fluiddiff","last_synced_at":"2026-05-17T17:12:08.070Z","repository":{"id":312451469,"uuid":"1047493827","full_name":"Arthur-Kitsuragi/FluidDiff","owner":"Arthur-Kitsuragi","description":"A Denoising Diffusion Model for Fluid Field Prediction","archived":false,"fork":false,"pushed_at":"2025-08-30T17:46:29.000Z","size":78759,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-08-30T19:09:53.108Z","etag":null,"topics":["diffusion-model","fluid-simulation","machine-learning","phiflow","tensorflow"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","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/Arthur-Kitsuragi.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,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-08-30T14:48:14.000Z","updated_at":"2025-08-30T17:46:32.000Z","dependencies_parsed_at":"2025-08-30T19:20:44.170Z","dependency_job_id":null,"html_url":"https://github.com/Arthur-Kitsuragi/FluidDiff","commit_stats":null,"previous_names":["arthur-kitsuragi/fluiddiff"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/Arthur-Kitsuragi/FluidDiff","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Arthur-Kitsuragi%2FFluidDiff","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Arthur-Kitsuragi%2FFluidDiff/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Arthur-Kitsuragi%2FFluidDiff/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Arthur-Kitsuragi%2FFluidDiff/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Arthur-Kitsuragi","download_url":"https://codeload.github.com/Arthur-Kitsuragi/FluidDiff/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Arthur-Kitsuragi%2FFluidDiff/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":273990205,"owners_count":25203290,"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":"2025-09-07T02:00:09.463Z","response_time":67,"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":["diffusion-model","fluid-simulation","machine-learning","phiflow","tensorflow"],"created_at":"2025-09-07T03:08:18.701Z","updated_at":"2026-05-17T17:12:08.029Z","avatar_url":"https://github.com/Arthur-Kitsuragi.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# FluidDiff\n\nThis repository contains an implementation of **FluidDiff**, a diffusion-based model for fluid dynamics.  \nThe project includes dataset generation, model training, and example usage.  \nOriginal paper: [link to the article](https://arxiv.org/pdf/2301.11661)  \n\n---\n\n## 📂 Repository Structure\n\n- **`FluidDiff.py`** — core implementation of the FluidDiff model (can be imported as a module).  \n- **`FluidDiff_dataset.ipynb`** — dataset generation using [PhiFlow](https://github.com/tum-pbs/PhiFlow).  \n- **`FluidDiff_train.ipynb`** — training notebook for the FluidDiff model.  \n- **`example.py`** — minimal example of model inference.  \n- **`example_data.npy`** — example dataset used for testing and inference.  \n- **`weights.40.weights.h5`** — trained model weights after 40 epochs.  \n\n---\n\n📌 Notes\n\nDesigned for fluid field simulations (e.g., velocity, pressure).\n\nTrained on synthetic data generated with PhiFlow.\n\nAchieved MAE = 0.1 after 40 epochs.\n\nTo predict future condition of fluid use `ddm.generate(n, diffusion_time, data)`, \nwhere `n` is amount of batches, `diffusion_time` is amount of diffusion steps during\nsampling, `data` is TensorFlow Tensor (shape = (N, 64, 64, 4), where N is number of pictures,\nthe first and the second channels should be Noise, the third one - ρ(0), the fourth one - \n64x64 matrix where each element = t/40 (t can be 1,...40; in fact it's simulation time))\n## Example Generation\n\nBelow is an example of FluidDiff denoising a fluid field:\n\n![Example Output](assets/example.png)\n\n## Architecture\n\n![Architecture](assets/Arch.PNG)","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Farthur-kitsuragi%2Ffluiddiff","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Farthur-kitsuragi%2Ffluiddiff","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Farthur-kitsuragi%2Ffluiddiff/lists"}