{"success":true,"data":{"author":"Macke Lab (mackelab)","bucketAccessType":"public","bucketName":"mackelab/sbi-for-connectomics-data","bucketPrefix":"","category":"Neuroscience","cloudProvider":"huggingface","createdAt":"2026-11-17T14:12:21.523Z","description":"Companion data for \"Simulation-based inference for efficient identification of generative models in connectomics\" from the Macke lab. The paper uses simulation-based inference (SBI, neural posterior estimation) to infer the parameters of computational wiring-rule models of the rat barrel cortex, a dense structural model of somatosensory cortex used in connectomics. About 2 to 3 GB across 219 files in three folders that mirror the paths the code expects: data/ (masks and feature tables for the structural barrel cortex model, the subcellular model and the cube model, plus presimulated prior and posterior samples for several wiring-rule parameterizations), results/ (amortized NPE posteriors, prior and posterior predictive samples, simulation-based calibration and cross-validation results, tutorial posteriors) and figures/ (source data and renders for the main and supplementary figures). Pickles are torch/sbi objects loaded with the consbi package and notebooks in github.com/mackelab/sbi-for-connectomics; tables are .txt/.tsv. Download into a clone of that repository with hf download. Released under the MIT License, matching the code repository; cite the bioRxiv preprint (2023.01.31.526269).","displayTitle":"SBI for Connectomics: wiring-rule simulations and posteriors for rat barrel cortex connectivity models","huggingFaceUrl":"https://huggingface.co/datasets/mackelab/sbi-for-connectomics-data","id":"37c7a88d-abe6-50f8-bc94-bc506ee30b20","license":"MIT","licenseEvidenceUrl":"https://huggingface.co/datasets/mackelab/sbi-for-connectomics-data","objectCountEstimate":221,"probeStatus":"ok","referenceUrl":"https://www.biorxiv.org/content/10.1101/2023.01.31.526269v1","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"mackelab/sbi-for-connectomics-data\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"mackelab/sbi-for-connectomics-data\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=mackelab, license:mit, downloads=419"],"format":"mixed","generatedAt":"2026-11-17T14:23:02.034Z","layout":{"example":"data/cube_model/features/features_set-6_float.tsv","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"# common_cubes_50\tcommon_cubes_25\tcommon_cubes_10\tcommon_cubes_5\tcommon_cubes_1\n7.0\t2.0\t0.0\t0.0\t0.0\n16.0\t17.0\t7.0\t3.0\t0.0\n14.0\t11.0\t4.0\t1.0\t0.0\n14.0\t20.0\t9.0\t2.0\t0.0\n15.0\t11.0\t4.0\t4.0\t2.0\n14.0\t10.0\t10.0\t6.0\t4.0\n8.0\t3.0\t0.0\t0.0\t0.0\n0.0\t0.0\t0.0\t0.0\t0.0\n1.0\t0.0\t0.0\t0.0\t0.0\n4.0\t3.0\t0.0\t0.0\t0.0\n2.0\t0.0\t0.0\t0.0\t0.0\n4.0\t1.0\t2.0\t0.0\t0.0\n3.0\t2.0\t1.0\t1.0\t0.0\n2.0\t1.0\t0.0\t0.0\t0.0\n4.0\t5.0\t2.0\t0.0\t0.0\n0.0\t0.0\t0.0\t0.0\t0.0\n0.0\t0.0\t0.0\t0.0\t0.0\n0.0\t0.0\t0.0\t0.0\t0.0\n1.0\t0.0\t0.0\t0.0\t0.0\n1.0\t0.0\t0.0\t0.0\t0.0\n1.0\t0.0\t0.0\t0.0\t0.0\n0","encoding":"utf8","key":"data/cube_model/features/features_set-6_float.tsv","truncatedAt":512}],"sampleObjects":["data/cube_model/depthplot/simulated_synapse_counts.p","data/cube_model/features/features_set-6_float.tsv","data/cube_model/features/features_set-6_int.tsv","data/cube_model/features/features_set-7_float.tsv","data/cube_model/features/features_set-7_int.tsv"],"sidecars":[{"bytes":"---\nlicense: mit\ntags:\n- connectomics\n- simulation-based-inference\n- neuroscience\n- computational-biology\npretty_name: SBI for Connectomics - Data and Results\n---\n\n# Dataset Card: SBI for Connectomics — Data and Results\n\nThis dataset contains the binary data, inference results, and figures accompanying the paper\n[Simulation-based inference for efficient identification of generative models in connectomics](https://www.biorxiv.org/content/10.1101/2023.01.31.526269v1).\n\nIt is the companion data repository for the code at\n[github.com/mackelab/sbi-for-connectomics](https://github.com/mackelab/sbi-for-connectomics), which uses\n[simulation-based inference (SBI)](http://simulation-based-inference.org) to infer parameters of computational\nwiring-rule models of the rat barrel cortex used in connectomics research.\n\nThese files were previously distributed via `git-lfs` in the code repository and have been moved here to keep the\ngit repository lightweight.\n\n## Dataset structure\n\n~3.0 GB across 219 binary files, organized into three top-level folders that mirror the paths expected by the\ncode and notebooks in the companion repository:\n\n```\ndata/\n├── structural_model/       # masks and feature tables for the structural (barrel cortex) model\n├── subcellular_features/   # masks and feature tables for the subcellular model\n├── cube_model/             # masks, feature tables, and depth-profile data for the cube model\n└── presimulated_dso_*.p    # presimulated prior/posterior samples for various wiring-rule parameterizations\n\nfigures/\n├── fig2_structural_model.pdf\n├── fig3/                   # main-text figure 3 (SBI results) source data/renders\n├── fig4-5/                  # main-text figures 4-5 (distance-rule results)\n└── supplementary_figures/   # supplementary figure renders\n\nresults/\n├── amortized_posterior_*.p          # amortized NPE posteriors\n├── npe_*.p                          # NPE posteriors trained on various wiring-rule datasets\n├── prior_predictive_*.p / posterior_predictive_*.p\n├── sbc_results_*.p                  # simulation-based calibration results\n├── cross_validation_results_*.p\n└── tutorial_posterior_*.p           # posteriors used in the tutorial notebooks\n```\n\nFile formats:\n- `.p` — Python pickle files (mostly `torch`/`sbi` posterior objects and sample arrays; see the companion\n  repository's `consbi` package and notebooks for how they're loaded)\n- `.txt` / `.tsv` — mask and feature tables for the structural/subcellular/cube models\n- `.pdf` / `.png` / `.eps` — rendered paper figures\n\nCode, notebooks, and small config/log files (plotting scripts, `.ipynb` notebooks, `.hydra` run configs, etc.)\nare **not** included here — those remain version-controlled in the\n[GitHub repository](https://github.com/mackelab/sbi-for-connectomics).\n\n## How to use\n\nDownload directly into a local clone of the [companion repository](https://github.com/mackelab/sbi-for-connectomics)\nso the folder layout matches what the code expects:\n\n```shell\npip install -U huggingface_hub\nhf download mackelab/sbi-for-connectomics-data --repo-type dataset --local-dir .\n```\n\nThis populates `data/`, `figures/`, and `results/` in the current directory.\n\n## Citation\n\nIf you use this data, please cite the associated paper:\n\n> Simulation-based inference for efficient identification of generative models in connectomics.\n> https://www.biorxiv.org/content/10.1101/2023.01.31.526269v1\n\n## License\n\nMIT License, matching the [companion code repository](https://github.com/mackelab/sbi-for-connectomics/blob/main/LICENSE.md).\n","encoding":"utf8","key":"README.md","name":"README.md","size":3636,"truncated":false}],"sourceObjectKey":"data/cube_model/features/features_set-6_float.tsv","starterCell":"import pandas as pd\ndf = pd.read_csv(\"hf://datasets/mackelab/sbi-for-connectomics-data/data/cube_model/features/features_set-6_float.tsv\", sep='\\t')\ndf.head()","starterSamplePath":"data/cube_model/features/features_set-6_float.tsv","totalObjects":221},"summary":"Companion data for \"Simulation-based inference for efficient identification of generative models in connectomics\" from the Macke lab.","title":"SBI for Connectomics: wiring-rule simulations and posteriors for rat barrel cortex connectivity models","updatedAt":"2026-11-17T20:20:19.821Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/mackelab/sbi-for-connectomics-data"}}