Dataset Viewer
The dataset viewer is not available for this subset.
Cannot get the split names for the config 'default' of the dataset.
Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/harbor/harbor.py", line 171, in _split_generators
                  raise DataFilesNotFoundError("No task.toml or instruction.md files found")
              datasets.exceptions.DataFilesNotFoundError: No task.toml or instruction.md files found
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ~~~~~~~~~~~~~~~~~~~~~~~^
                      path=dataset,
                      ^^^^^^^^^^^^^
                      config_name=config,
                      ^^^^^^^^^^^^^^^^^^^
                      token=hf_token,
                      ^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
                  info = get_dataset_config_info(
                      path,
                  ...<6 lines>...
                      **config_kwargs,
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Berlin 3D Mesh 2025 – 3D Tiles archive

An archival copy of the textured, photogrammetric 3D mesh of Berlin (state 2025) in OGC 3D Tiles 1.0 format (.b3dm with glTF 2 geometry and JPEG textures).

The official Berlin 3D download portal (https://www.businesslocationcenter.de/berlin3d-downloadportal/) has shown a maintenance page since about October 2026. This archive keeps the data available.

Status: upload in progress. Shards are added as they finish; see "Completeness" below.

Source and attribution

  • Provider / attribution: Berlin Partner für Wirtschaft und Technologie GmbH (dataset "3D-Meshmodell Berlin", listed at https://daten.berlin.de/datensaetze/3d-meshmodell-berlin).
  • Mesh basis: multi-perspective aerial photo flight 2025.
  • Copied from: https://berlin-wab.virtualcitymap.de/datasource-data/baselayer/mesh_2025/tileset.json (3D Tiles stream used by the Berlin Wirtschaftsatlas).
  • Crawled: 2026-11-15.
  • Permission: published with permission to redistribute, obtained 2026-11-15. See LICENSE.md.

When you use the data, credit: "3D-Meshmodell Berlin 2025 © Berlin Partner für Wirtschaft und Technologie GmbH".

Contents

Tileset JSONs 109,086
Tiles (.b3dm) 742,387
Total size ≈ 284 GB (estimate, exact figure in STATS.md once complete)
Coverage (root bbox, WGS84) 12.83–13.81 °E, 52.16–52.73 °N (all of Berlin)
Detail levels L0 (finest, ≈ 1 m geometric error) … L17 (coarsest)

Tiles per level: L0 552,661 · L1 144,282 · L2 32,655 · L3 9,242 · L4 2,205 · L5 764 · L6 198 · L7 198 · L8–L17 182.

Repository layout

tilesets.tar                 all tileset JSONs incl. root tileset.json + _tiles.txt (full tile list)
tiles/shard_NNNN.tar         ~4 GB each, .b3dm tiles at their original relative paths
manifests/shard_NNNN.sha256  SHA-256 of every tile in that shard (sha256sum format)

Restore the full tileset

hf download tifa365/berlin-mesh-2025 --repo-type dataset --local-dir berlin-mesh-2025
cd berlin-mesh-2025
mkdir tileset
tar -xf tilesets.tar -C tileset
for t in tiles/*.tar; do tar -xf "$t" -C tileset; done
cat manifests/*.sha256 > tileset/MANIFEST.sha256
(cd tileset && sha256sum -c --quiet MANIFEST.sha256 && echo "all tiles OK")

tileset/tileset.json is then a complete 3D Tiles dataset. Only one district? Each shard holds neighbouring tiles; tilesets.tar alone (≈ 700 MB unpacked) lets you see which tiles exist.

View and convert

  • CesiumJS: serve the folder over HTTP and load it with await Cesium.Cesium3DTileset.fromUrl("http://localhost:8000/tileset.json").
  • To glTF/GLB: npx 3d-tiles-tools b3dmToGlb -i tile.b3dm -o tile.glb (https://github.com/CesiumGS/3d-tiles-tools).
  • Tile positions are in ECEF via each tileset node's transform.

How this archive was made

  1. Crawled the whole tileset tree from the root tileset.json (no directory listing exists). 9 sub-tilesets returned HTTP 502 and were re-crawled successfully: 0 open errors.
  2. Downloaded tiles in ~4 GB batches; every tile was checked (b3dm magic, declared length, glTF payload) before packing, and every uploaded shard's remote size was verified.

Completeness

Updated when the upload finishes: number of shards, tiles, bytes, and any tiles that could not be downloaded (expected: none).

Downloads last month
-