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/webdataset/webdataset.py", line 84, in _split_generators
                  raise ValueError(
                  ...<2 lines>...
                  )
              ValueError: The TAR archives of the dataset should be in WebDataset format, but the files in the archive don't share the same prefix or the same types.
              
              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.

Diogenes Legacy EXR Depth Experiments Archive v01

The project has closed. This owner-authorized archive preserves 178,348 original EXR frames and 78 original JSONL indices / 35,501,732,506 source bytes across 120 historical capture directories. The complete reviewed collection is published and verified in 286 TAR shards. MANIFEST.jsonl records the actual SHA-256 and byte size of every member; ARCHIVE_INDEX.json links each immutable verification revision and TAR hash. All corresponding R2 source objects were removed only after pinned Hub LFS metadata and anonymous-download checks passed.

Scope and format

These are archived capture/debugging attempts from the Fallout 4, Euro Truck Simulator 2 and Cyberpunk 2077 development lineage, not a newly certified training dataset. Paths retain original issue/session/attempt identifiers so related evidence can be joined without inferring a new success status. Some parent folders group several games or copied artifacts: a folder name alone is not reliable game attribution.

A representative header from every one of the 120 frame directories was parsed: each is 1280×720, single-channel Z, half-precision EXR, with only standard numeric attributes and generic depth / scanlineimage labels. This establishes the sampled file representation, not consistent depth units, calibrated geometry, projection validity or full-frame health. Earlier attempts can contain clear/invalid buffers, duplicates and failures. Original storage, projection, health, buffer-format, DSV-resource and latency fields in the associated index must be consulted; no metre-depth or training eligibility claim is inferred from an EXR extension.

The 78 complete source indices contain 120,999 rows referencing 114,996 distinct stored EXRs. Repeated rows remain intact. Every referenced path resolves within that capture branch. Another 63,352 EXRs have no corresponding complete index within this initial archive scope: 59,752 offloaded frames and 3,600 unpacked delivery frames. Their original relative paths are retained, but absent timing/calibration must not be invented. The 59,752 offload destinations replace one machine-hostname component with capture-host; bytes are unchanged. Adjacent RGB, audio, pose streams, ZIP packages and broader diagnostics are outside this initial frame manifest and must not be assumed available here.

Review and shards

All 78 included index objects were read fully for schema, strings and privacy; their frame references were checked against the source inventory. EXR header review sampled the middle frame of each directory and found no custom text identity fields. It did not exhaustively decode every image or certify the numerical content.

Shards contain up to 900 intact consecutive files from one directory, with its index in the first shard where available. No frame is split. Each tar ends with a per-member SHA-256/size manifest. Collect the branch's shards to reconstruct original paths; read the index beside its raw/ subdirectory. Per-source SHA-256, fixed Hugging Face LFS size/hash and anonymous-download checks are completed during transfer before source deletion.

Public availability does not change historical non-trainable, failed-QC or missing-certification states. See NOTICE.md. Capture source: Diogenes-Archive.

Downloads last month
212