Datasets:

Yangyihui
/

3828612ec258e8fc

License:

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 80, 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.

Internal dataset artifact

Internal identifier: 3828612ec258e8fc

Transfer artifact for an internal experiment. Not licensed for reuse, redistribution, mirroring, or third-party model training.

package	clips	scenes	shards
`train/`	2600	650	21
`extra_legacy_rescued/train/`	396	276	4

Video: 512x288, 4 FPS, h264.
Action: body-frame SE(2) deltas at dt = 0.25 s, trajectory.npz:action_delta_body.
Every clip carries one instruction only, in the canonical go to the ... form.
Clip lineage tag: episode.json:data_category / meta.jsonl:source_id.

`data_category`	clips
`far_v070`	429
`legacy_rerender`	1516
`midfar`	655

extra_legacy_rescued is a separate package on purpose: it is a later increment and must not be mixed into the main set when comparing runs.

Use

python tools/materialize_scan2d_hf_split.py \
    --source /path/to/3828612ec258e8fc --split train --output /path/to/dataset
python tools/materialize_scan2d_hf_split.py \
    --source /path/to/3828612ec258e8fc/extra_legacy_rescued --split train \
    --output /path/to/dataset_extra

Materialization verifies every shard checksum and tar member list, then copies the index files out of train/:

awm_meta.jsonl - AWM training index (2600 rows)
train.txt / val.txt - 2460 / 140 clips, scene-disjoint (32 validation scenes, ratio 0.05, seed 42)
val_scenes.txt - the validation scene list, so the same split can be reproduced for any subset of this release
meta.jsonl / all.txt - release index, one row per clip
text_variants_diverse.jsonl - optional alternative phrasing per clip; the clip files themselves stay canonical, so ignoring this file leaves the canonical-only view intact
ambiguity_verdicts.jsonl - per-clip audit provenance

train/shards.json records every tar checksum.

Downloads last month: 20

Total file size:

2.71 GB

Internal dataset artifact

Contents

Use