{"success":true,"data":{"author":"Blumenstiel, Ait Ali Braham, Albrecht, Maurogiovanni, Fraccaro (IBM Research, DLR, TUM; Embed2Scale)","bucketAccessType":"public","bucketName":"embed2scale/SSL4EO-S12-v1.1","bucketPrefix":"","category":"Earth Observation & Geospatial","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:15.513Z","description":"Updated release of SSL4EO-S12, the self-supervised learning dataset for Earth observation built from Sentinel-1 and Sentinel-2. It covers 246,144 locations worldwide, each sampled at four seasonal timestamps, with co-registered modalities S2L1C, S2L2A, S2RGB, S1GRD (VV and VH backscatter), land use and land cover (LULC, augmented with cloud and snow masks), a digital elevation model (DEM) and NDVI. Version 1.1 fixes alignment and metadata issues in the 2022 original, and it is one of the two sources of the TerraMesh pretraining corpus. Samples are Zarr v2 zip files with aligned metadata (cloud mask, centre coordinates, CRS, product ids), grouped into tar shards for loading with webdataset; the listing (measured 2026-11-14) shows train/ and val/ folders with one subfolder per modality plus train_metadata.parquet and val_metadata.parquet holding per-sample timestamps and cloud cover. The card gives about 2.3 TB for the full download and xarray or zarr opens single samples. A Zarr-chunk variant lives in embed2scale/SSL4EO-S12-v1.1-Zarr. The authors release the dataset under CC-BY-4.0 (code Apache-2.0); cite arXiv:2503.00168 and the original SSL4EO-S12 paper (Wang et al., IEEE GRSM 2023).","displayTitle":"SSL4EO-S12 v1.1: Multimodal, Multiseasonal Sentinel-1/2 Pretraining Dataset","id":"0d77cf8f-a0c1-561e-98e4-808e255dad9e","license":"CC-BY-4.0","licenseEvidenceUrl":"https://github.com/DLR-MF-DAS/SSL4EO-S12-v1.1","objectCountEstimate":3378,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2503.00168","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"embed2scale/SSL4EO-S12-v1.1\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"embed2scale/SSL4EO-S12-v1.1\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=embed2scale, license:cc-by-4.0, size_categories:100K<n<1M, format:parquet, downloads=7687"],"format":"mixed","generatedAt":"2026-11-14T14:15:26.035Z","layout":{"example":"train_metadata.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"50415231150415ccdf0115f8254c15ba071500120000e66f501a00000073736c34656f7331325f73686172645f30010110312e746172621e000032721e000033721e000034721e000035721e000036721e000037721e000038721e0000396e1e00043130721e00722c010031722c010031722c010031722c010031722c010031722c010031722c010031722c010031722c010032722c010032722c010032722c010032722c010032722c010032722c010032722c010032722c010032722c010032722c010033722c010033722c010033722c010033722c010033722c010033722c010033722c010033722c010033722c010033722c010034722c010034722c010034722c010034722c010034722c010034722c010034722c010034722c010034722c010034722c010035722c010035722c010035722c010035722c010035722c010035722c010035722c010035722c010035722c010035722c010036722c010036722c010036722c010036722c010036722c010036722c010036722c010036722c010036722c010036722c010037722c010037722c010037722c010037722c010037722c010037722c010037722c010037722c010037722c010037722c010038722c010038722c010038722c010038722c010038722c010038722c010038722c010038722c010038722c010038722c010039722c010039722c010039722c0100","encoding":"hex","key":"train_metadata.parquet","truncatedAt":512}],"sampleObjects":["train/DEM/ssl4eos12_shard_000001.tar","train/DEM/ssl4eos12_shard_000002.tar","train/DEM/ssl4eos12_shard_000003.tar","train/DEM/ssl4eos12_shard_000004.tar","train/DEM/ssl4eos12_shard_000005.tar"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntags:\n- Earth Observation\n- Pretraining\n- Multimodal\nsize_categories:\n- 100K<n<1M\ntask_categories:\n- feature-extraction\nlibrary_name: webdataset\nviewer: true\nconfigs:\n- config_name: default\n  data_files:\n  - split: train\n    path: \"train_metadata.parquet\"\n  - split: validation\n    path: \"val_metadata.parquet\"\n---\n\n[![arXiv](https://img.shields.io/badge/arXiv-2503.00168-b31b1b?logo=arxiv)](https://arxiv.org/abs/2503.00168)\n[![GitHub](https://img.shields.io/badge/GitHub-SSL4EOS12v1.1-AAAAAA?logo=github)](https://github.com/DLR-MF-DAS/SSL4EO-S12-v1.1)\n\n# SSL4EO-S12-v1.1\n\n> Attention: The Zarr Chunk file version of SSL4EO-S12-v1.1 was moved to [embed2scale/SSL4EO-S12-v1.1-Zarr](https://huggingface.co/datasets/embed2scale/SSL4EO-S12-v1.1-Zarr). This repository contains data that can be used with `webdataset`.\n\nThe dataset includes 246,144 locations with four timestamps each from the modalities S2L1C, S2L2A, S2RGB, S1GRD, LULC, DEM, and NDVI.\nWe refer to our [technical report](https://arxiv.org/abs/2503.00168) for details.\n\nThe samples are stored in as Zarr Zip files (zarr version 2) with the metadata directly aligned as additional data variables.\nThe samples are randomly grouped in `tar` shards and can be loaded easily via `webdataset`.\nBecause we want to keep the loading of specific modalities flexible, we differ from the standard and save each modality in separate folders. \nTherefore, we share custom code for data loading via webdataset in our [GitHub repository](https://github.com/DLR-MF-DAS/SSL4EO-S12-v1.1). \n\nSentinel-2 and Sentinel-1 time series examples with four seasonal images:\n\n![ssl4eos12_timeseries.png](https://raw.githubusercontent.com/DLR-MF-DAS/SSL4EO-S12-v1.1/refs/heads/main/assets/ssl4eos12_timeseries.png)\n\nModality examples in SSL4EO-S12 v1.1:\n\n![ssl4eoS12_modalities.png](https://raw.githubusercontent.com/DLR-MF-DAS/SSL4EO-S12-v1.1/refs/heads/main/assets/ssl4eoS12_modalities.png)\n\nSentinel-2 L2A example, loaded as xarray dataset:\n```text\n<xarray.Dataset> Size: 7MB\nDimensions:      (time: 4, band: 12, y: 264, x: 264)\nCoordinates:\n  * time         (time) datetime64[ns] 32B 2020-11-16T02:30:01 ... 2021-08-08...\n  * band         (band) <U3 144B 'B01' 'B02' 'B03' 'B04' ... 'B09' 'B11' 'B12'\n  * y            (y) float64 2kB 4.728e+06 4.728e+06 ... 4.725e+06 4.725e+06\n  * x            (x) float64 2kB 7.149e+05 7.149e+05 ... 7.175e+05 7.175e+05\n    sample       <U7 28B '0080717'\n    spatial_ref  int64 8B 0\nData variables:\n    bands        (time, band, y, x) int16 7MB 463 457 451 445 ... 686 663 635\n    center_lat   float64 8B 42.66\n    center_lon   float64 8B 125.6\n    cloud_mask   (time, y, x) uint8 279kB 0 0 0 0 0 0 0 0 0 ... 0 0 0 0 0 0 0 0\n    crs          int64 8B 32651\n    file_id      (time) <U38 608B '20201116T023001_20201116T022955_T51TYH' .....\n    sample_id    (time) <U9 144B '0080717_0' '0080717_1' '0080717_2' '0080717_3'\n```\n\nSentinel-1 GRD example, loaded as xarray dataset:\n```text\nDimensions:      (time: 4, band: 2, y: 264, x: 264)\nCoordinates:\n  * time         (time) datetime64[ns] 32B 2020-11-27T21:46:46 ... 2021-08-30...\n  * band         (band) <U2 16B 'vv' 'vh'\n  * y            (y) float64 2kB 4.728e+06 4.728e+06 ... 4.725e+06 4.725e+06\n  * x            (x) float64 2kB 7.149e+05 7.149e+05 ... 7.175e+05 7.175e+05\n    sample       <U7 28B '0080717'\n    spatial_ref  int64 8B 0\nData variables:\n    bands        (time, band, y, x) float16 1MB -15.13 -17.44 ... -18.64 -19.33\n    center_lat   float64 8B 42.66\n    center_lon   float64 8B 125.6\n    crs          int64 8B 32651\n    file_id      (time) <U67 1kB 'S1B_IW_GRDH_1SDV_20201127T214646_20201127T2...\n    sample_id    (time) <U9 144B '0080717_0' '0080717_1' '0080717_2' '0080717_3'\n```\n\n## Download\n\nYou can download the dataset via the Hugging Face CLI (`pip install huggingface_hub`). Please note that the full dataset requires 2.3TB of storage.\n\n```shell\nhf download embed2scale/SSL4EO-S12-v1.1 --repo-type dataset --local-dir data/SSL4EOS12\n```\n\nIf you like to download only a subset of the data, you can specify it with `--include`.\n","encoding":"utf8","key":"README.md","name":"README.md","size":7250,"truncated":true}],"sourceObjectKey":"train_metadata.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"embed2scale/SSL4EO-S12-v1.1\")\nprint(ds)","starterSamplePath":"train/DEM/ssl4eos12_shard_000001.tar","totalObjects":3378},"summary":"Updated release of SSL4EO-S12, the self-supervised learning dataset for Earth observation built from Sentinel-1 and Sentinel-2.","title":"SSL4EO-S12 v1.1: Multimodal, Multiseasonal Sentinel-1/2 Pretraining Dataset","updatedAt":"2026-11-14T17:53:48.851Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/embed2scale/SSL4EO-S12-v1.1"}}