{"success":true,"data":{"author":"Polymathic AI (The Well); Rudy Morel, Flatiron Institute CCM","bucketAccessType":"public","bucketName":"polymathic-ai/shear_flow","bucketPrefix":"","category":"Physics","cloudProvider":"huggingface","createdAt":"2026-11-14T01:01:19.189Z","description":"2D periodic shear flow governed by the incompressible Navier-Stokes equations with an advected passive tracer, solved spectrally with Dedalus. Each simulation has 200 time steps of 128 by 256 images over t from 0 to 20 on a doubly periodic domain. Fields are tracer, pressure and velocity. The datasheet lists 1,120 simulations: 28 PDE parameter pairs (Reynolds 1e4 to 5e5, Schmidt 0.1 to 10) times 40 initial conditions built from tanh shear layers with sinusoidal perturbations. The publishers warn that this Hugging Face copy is v1.0, in which about 80 percent of validation and test trajectories duplicate training trajectories; a corrected v2.0 with disjoint initial conditions is available from Flatiron through the-well-download. Listing data/ (measured 2026-11-13) shows train/ (28 files, 376 GB, one per Reynolds-Schmidt pair), valid/ and test/ (28 files, 47.5 GB each); about 471 GB in the repository. Files are HDF5 in the common The Well format; open them with the_well package (WellDataModule with base path hf://datasets/polymathic-ai/) or directly with h5py. Part of The Well (Ohana et al., NeurIPS 2024 Datasets and Benchmarks, arXiv:2412.00568); released under CC-BY-4.0, cite The Well and the simulation's source papers.","displayTitle":"The Well: Periodic Shear Flow (2D Incompressible Navier-Stokes)","id":"5fbaecc3-72e6-5fe1-9260-593bb6da2402","license":"CC-BY-4.0","licenseEvidenceUrl":"https://arxiv.org/html/2412.00568","objectCountEstimate":91,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2412.00568","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"polymathic-ai/shear_flow\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"polymathic-ai/shear_flow\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=polymathic-ai, license:cc-by-4.0, downloads=4969"],"format":"mixed","generatedAt":"2026-11-14T01:01:25.083Z","layout":{"example":"data/test/shear_flow_Reynolds_1e4_Schmidt_1e-1.hdf5","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/test/shear_flow_Reynolds_1e4_Schmidt_1e-1.hdf5","truncatedAt":512}],"sampleObjects":["data/test/shear_flow_Reynolds_1e4_Schmidt_1e-1.hdf5","data/test/shear_flow_Reynolds_1e4_Schmidt_1e0.hdf5","data/test/shear_flow_Reynolds_1e4_Schmidt_1e1.hdf5","data/test/shear_flow_Reynolds_1e4_Schmidt_2e-1.hdf5","data/test/shear_flow_Reynolds_1e4_Schmidt_2e0.hdf5"],"sidecars":[{"bytes":"---\nlanguage:\n- en\nlicense: cc-by-4.0\ntags:\n- physics\n\ntask_categories:\n- time-series-forecasting\n- other\ntask_ids:\n- multivariate-time-series-forecasting\n---\n\n# How To Load from HuggingFace Hub\n\n1. Be sure to have `the_well` installed (`pip install the_well`)\n2. Use the `WellDataModule` to retrieve data as follows:\n\n```python\nfrom the_well.data import WellDataModule\n\n# The following line may take a couple of minutes to instantiate the datamodule\ndatamodule = WellDataModule(\n    \"hf://datasets/polymathic-ai/\",\n    \"shear_flow\",\n)\ntrain_dataloader = datamodule.train_dataloader()\n\nfor batch in dataloader:\n    # Process training batch\n    ...\n```\n\n# Periodic shear flow\n\n> **Warning: this Hugging Face copy is v1.0 of the dataset, which has an error in its split.** About 80% of its validation and test trajectories are exact copies of training trajectories (same PDE parameters and same initial condition), so validation and test results obtained on it are not reliable.\n>\n> A corrected version (v2.0), with train, validation and test built from disjoint initial conditions, is available from Flatiron with:\n>\n> ```\n> the-well-download --base-path path/to/base --dataset shear_flow\n> ```\n>\n> If you already downloaded `shear_flow` before October 2026, delete your local copy first: the download resumes existing files and would otherwise keep the old ones. This Hugging Face copy will be updated to v2.0 soon.\n\n**One line description of the data:** 2D periodic incompressible shear flow.\n\n**Longer description of the data:**\nShear flow are a type of fluid characterized by the continuous deformation of adjacent fluid layers sliding past each other with different velocities. This phenomenon is commonly observed in various natural and engineered systems, such as rivers, atmospheric boundary layers, and industrial processes involving fluid transport.\nThe dataset explores a 2D periodic shearflow governed by incompressible Navier-Stokes equation.\n\n**Associated paper**: [Paper 1](https://www.sciencedirect.com/book/9780124059351/fluid-mechanics), [Paper 2](https://journals.aps.org/prfluids/abstract/10.1103/PhysRevFluids.6.100504), [Paper 3](https://journals.aps.org/prfluids/abstract/10.1103/PhysRevFluids.9.014202).\n\n**Data generated by**: [Rudy Morel](https://rudymorel.github.io/), CCM, Flatiron Institute.\n\n**Code or software used to generate the data**: [Github repository](https://github.com/RudyMorel/the-well-rbc-sf), based on the software [Dedalus](https://dedalus-project.readthedocs.io/en/latest/pages/examples/ivp_2d_shear_flow.html).\n\n**Equation**:\n\nWhile we solve equations in the frequency domain, the original time-domain problem is\n\n$$\n\\begin{align*}\n\\frac{\\partial u}{\\partial t} + \\nabla p - \\nu \\Delta u & = -u\\cdot\\nabla u\\,,\n\\\\\n\\frac{\\partial s}{\\partial t} - D \\Delta s & = -u \\cdot\\nabla s\\,,\n\\end{align*}\n$$\n\nwhere \\\\(\\Delta = \\nabla \\cdot \\nabla\\\\) is the spatial Laplacian, \\\\(u = (u_x,u_y)\\\\) is the (horizontal and vertical) velocity, \\\\(s\\\\) is the tracer, and \\\\(p\\\\) is the pressure,\nwith the additional constraints \\\\(\\int p = 0\\\\) (pressure gauge).\n\nThese PDEs are parameterized by the Reynolds and Schmidt numbers through \\\\(\\nu\\\\) and \\\\(D\\\\).\n\n$$\n\\begin{align*}\n\\text{(viscosity)} ~~~~~~~ \\nu & = 1 / \\text{Reynolds}\n\\\\\n\\text{(diffusivity)} ~~~~~~~ D & = \\nu / \\text{Schmidt}\n\\end{align*}\n$$\n\n![Gif](https://users.flatironinstitute.org/~polymathic/data/the_well/datasets/shear_flow/gif/tracer_normalized.gif)\n\n| Dataset    | FNO | TFNO  | Unet | CNextU-net\n|:-:|:-:|:-:|:-:|:-:|\n| `shear_flow`  | 0.1567 | \\\\(\\mathbf{0.1348}\\\\) |0.5910|0.2037|\n\nTable: VRMSE metrics on test sets (lower is better). Best results are shown in bold. VRMSE is scaled such that predicting the mean value of the target field results in a score of 1.\n\n## About the data\n\n**Dimension of discretized data:** 200 time-steps of 128 \\\\(\\times\\\\) 256 images.\n\n**Fields available in the data:** tracer (scalar field), pressure (scalar field), velocity (vector field).\n\n**Number of simulations:** 1120 (28 PDE parameters \\\\(\\times\\\\) 40 initial conditions).\n\n**Size of the ensemble","encoding":"utf8","key":"README.md","name":"README.md","size":6611,"truncated":true}],"sourceObjectKey":"data/test/shear_flow_Reynolds_1e4_Schmidt_1e-1.hdf5","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"polymathic-ai/shear_flow\", filename=\"data/test/shear_flow_Reynolds_1e4_Schmidt_1e-1.hdf5\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"data/test/shear_flow_Reynolds_1e4_Schmidt_1e-1.hdf5","totalObjects":91},"summary":"2D periodic shear flow governed by the incompressible Navier-Stokes equations with an advected passive tracer, solved spectrally with Dedalus.","title":"The Well: Periodic Shear Flow (2D Incompressible Navier-Stokes)","updatedAt":"2026-11-14T02:45:00.037Z","visibility":"public","websiteUrl":"https://polymathic-ai.org/the_well/"}}