The dataset viewer is not available for this subset.
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.
Didi Mock App 4,000 Candidate Trajectories v1
This public package contains 4,000 deterministic candidate trajectories from two independent, immutable MobileGym-Harmony Didi Mock App batches. The measured package source contains 66,331 action rows and 66,331 screenshots captured immediately before those actions.
Covered synthetic workflows
- Destination selection and multi-vehicle calling.
- Waypoint selection, price comparison, and lowest-price choice.
- Strict timeout cancellation of the first local synthetic order, switching to a different vehicle type, and local matching of the second synthetic order.
The package is not a single homogeneous batch. batch_20260813_two_flow contains 2,000 candidates covering workflows 1 and 2. batch_20260817_timeout_cancel_switch contains 2,000 candidates covering workflow 3. Batch-specific trajectory schemas are preserved separately.
Synthetic-data boundary
All locations, orders, waits, cancellations, vehicles, plates, driver-like records, prices, and ETA values are local synthetic Mock App state. This is not real Didi collection and contains no real users, real locations, real orders, real drivers, real plates, payments, driver contact, location sharing, or credentials.
This release does not contain original sampling packages, restricted visual-oracle material, internal contracts, logs, progress records, recovery logs, browser logs, or local filesystem paths.
Verification and limitations
- Scope:
candidate_experiment_only - Status:
pass_with_non_promotion_boundary - Highest complete verification:
V4 - Didi readiness:
not_ready - Formal Loop 2:
T0_only / blocked_at_T0 - Batch A manual visual review:
not_performed_user_opt_out - Batch B manual visual review:
not_performed
A 100% deterministic validation pass rate must not be interpreted as a real Agent, real model, or real Didi success rate. These are not real-model rollouts, formal production trajectories, or data proven suitable for training or evaluation. V5 was not performed. Anonymous public download smoke testing is not V6 business or user acceptance.
Package layout
data/<batch_id>/: 250-trajectorytar.gzshards, with batch boundaries preserved.checksums/<batch_id>/: per-shard content SHA-256 manifests.metadata/<batch_id>/: frozen public-safe task and scenario metadata.schema/: one shared action space plus separate trajectory schemas for each batch.summaries/<batch_id>/: sanitized aggregate quality summaries.RELEASE-MANIFEST.json: release identity, batch lineage, measured counts, archive inventory, and per-trajectory mapping.SHA256SUMS: SHA-256 for the exact publication files, excludingSHA256SUMSitself.
License and rights boundary
This repository uses Hugging Face metadata value license: other. No standardized open-source or open-data license is declared. Public access does not automatically grant additional rights to copy, modify, redistribute, sublicense, or commercially use the materials. See LICENSE for the controlling notice.
Didi/滴滴 names and trademarks belong to their respective rights holders. This dataset is not affiliated with, endorsed by, sponsored by, or an official collaboration with Didi or any related entity.
- Downloads last month
- 75