The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: ValueError
Message: Expected object or value
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 290, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to string in row 0
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 304, in _generate_tables
batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
examples = [ujson_loads(line) for line in original_batch.splitlines()]
~~~~~~~~~~~^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
return pd.io.json.ujson_loads(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
ValueError: Expected object or valueNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
S1 + MLLM4D-20% training transfer bundle
This repository contains data and pretrained weights only; it does not contain training code.
Frozen training inputs
- S1:
llava_hound_64k,spar_234k, 298,027 rows. - MLLM4D:
mllm4d_2m%20, Python RNG seed 0, 399,999 sampled rows, 50,030 unique videos. - Base model:
Qwen/Qwen3-VL-8B-Instruct, revision0c351dd01ed87e9c1b53cbc748cba10e6187ff3b. - Geometry encoder:
edgarsucar/vdpm, revisionf38b3d8cb26543eda8954c865492f3a90bb3c7e5, filemodel.pt.
The exact MLLM4D-20% training population has membership SHA-256
7cbe158fe234a876400b240392ed8c2b0e7e67bbf8817c627adf51621295056a
and post-shuffle order SHA-256
8d265636d75497c2c233676f7f47585315aca74769da649c14c8bb4b9ef3c654.
Restore
All payload archives are independently extractable except
s1_spar_234k_media_03.tar, which is transported as four byte-exact parts to
avoid a large-file metadata timeout. Reassemble and verify it first:
cat s1_spar_234k_media_03.tar.part{00..03} > s1_spar_234k_media_03.tar
sha256sum -c s1_spar_234k_media_03.original.sha256
Then, from the intended project root:
sha256sum -c MANIFEST.sha256
for archive in /path/to/download/*.tar; do
tar -xf "$archive"
done
Extraction reconstructs the expected ckpts/ and data/ paths. In particular, the LLaVA-Hound media archive materializes the legacy paths required by the original S1 annotation under data/media/llava_hound/frames/.
annotations_s1_mllm4d20.tar contains the original full MLLM4D annotation so that mllm4d_2m%20 can reproduce the frozen selection. mllm4d20_seed0_selection_receipts.tar additionally contains the exact pre-shuffle and post-shuffle 399,999-row populations and audit inventories.
Upstream attribution
- Qwen3-VL-8B-Instruct: Qwen team, Apache-2.0.
- VDPM: Edgar Sucar et al., CC-BY-NC-4.0 upstream.
- SPAR-7M-RGBD: Fudan ZVG SPAR project.
- LLaVA-Hound media: ShareGPTVideo/LLaVA-Hound source material.
- MLLM-4D: GVCLab MLLM-4D project.
See MANIFEST.sha256 and BUNDLE_RECEIPT.json for exact archive hashes, sizes, populations, and source revisions.
- Downloads last month
- 75