Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
blurry: list<item: string>
child 0, item: string
sequence: string
sharp: list<item: string>
child 0, item: string
teacher: list<item: string>
child 0, item: string
teacher_kind: string
schema: string
checkpoint_sha256: string
sequence_count: int64
frame_count: int64
output_manifest: string
input_manifest_sha256: string
output_manifest_sha256: string
frames: list<item: struct<sequence_index: int64, frame_index: int64, blurry: string, blurry_sha256: string, (... 78 chars omitted)
child 0, item: struct<sequence_index: int64, frame_index: int64, blurry: string, blurry_sha256: string, sharp: stri (... 66 chars omitted)
child 0, sequence_index: int64
child 1, frame_index: int64
child 2, blurry: string
child 3, blurry_sha256: string
child 4, sharp: string
child 5, sharp_sha256: string
child 6, teacher: string
child 7, teacher_sha256: string
checkpoint: string
input_manifest: string
to
{'schema': Value('string'), 'input_manifest': Value('string'), 'input_manifest_sha256': Value('string'), 'checkpoint': Value('string'), 'checkpoint_sha256': Value('string'), 'output_manifest': Value('string'), 'output_manifest_sha256': Value('string'), 'sequence_count': Value('int64'), 'frame_count': Value('int64'), 'frames': List({'sequence_index': Value('int64'), 'frame_index': Value('int64'), 'blurry': Value('string'), 'blurry_sha256': Value('string'), 'sharp': Value('string'), 'sharp_sha256': Value('string'), 'teacher': Value('string'), 'teacher_sha256': Value('string')})}
because column names don't match
Traceback: 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 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
blurry: list<item: string>
child 0, item: string
sequence: string
sharp: list<item: string>
child 0, item: string
teacher: list<item: string>
child 0, item: string
teacher_kind: string
schema: string
checkpoint_sha256: string
sequence_count: int64
frame_count: int64
output_manifest: string
input_manifest_sha256: string
output_manifest_sha256: string
frames: list<item: struct<sequence_index: int64, frame_index: int64, blurry: string, blurry_sha256: string, (... 78 chars omitted)
child 0, item: struct<sequence_index: int64, frame_index: int64, blurry: string, blurry_sha256: string, sharp: stri (... 66 chars omitted)
child 0, sequence_index: int64
child 1, frame_index: int64
child 2, blurry: string
child 3, blurry_sha256: string
child 4, sharp: string
child 5, sharp_sha256: string
child 6, teacher: string
child 7, teacher_sha256: string
checkpoint: string
input_manifest: string
to
{'schema': Value('string'), 'input_manifest': Value('string'), 'input_manifest_sha256': Value('string'), 'checkpoint': Value('string'), 'checkpoint_sha256': Value('string'), 'output_manifest': Value('string'), 'output_manifest_sha256': Value('string'), 'sequence_count': Value('int64'), 'frame_count': Value('int64'), 'frames': List({'sequence_index': Value('int64'), 'frame_index': Value('int64'), 'blurry': Value('string'), 'blurry_sha256': Value('string'), 'sharp': Value('string'), 'sharp_sha256': Value('string'), 'teacher': Value('string'), 'teacher_sha256': Value('string')})}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Local space recovery archive — 2026-09-15 r1
This private archive records the inputs removed from /home/szha0669 and
/srv/szha0669 to recover at least 400 GiB of local disk space. It does not
claim that third-party datasets or model weights are owned by this archive.
Those bytes are not re-uploaded: their original repositories, immutable
revisions, file identities, preparation scripts, and local inventories are
recorded instead.
Final result
The cleanup completed successfully. Filesystem accounting measured
430,787,661,824 newly available bytes: 401.20 GiB / 430.79 GB. / gained
102.43 GiB and /srv gained 298.77 GiB. All 45 explicit targets were absent
after cleanup; no unique experiment-result tree was part of the target list.
What is protected
- Unique experiment results under
output_ablation/,turtle_finetune/, andexperiments/are retained locally. - The active BlurAdaptGS Python environment, PRISM3D preparation tools, portable configs, and compact input tree are retained locally.
- The complete PRISM3D-E8 50K result is independently available in
qizhangslam/blur-slam-exp-best-preservation-s1and passed a 604-file local/remote content audit before its local publication staging was removed. - The original user-provided r2 pipeline tarball and checksum remain in the Codex attachments directory.
Immutable external sources
| Content | Repository | Type | Revision |
|---|---|---|---|
| REDS archives | snah/REDS |
dataset | 62dc25d16e6f43d2214f1b365023abda86f7a0ae |
| GoPro Large archive | snah/GOPRO_Large |
dataset | 592978466ae510d2734b199cad2fc79a346bda1c |
| Unblur-SLAM curated data | qizhangslam/Unblur_slam_traning_dataset |
dataset | 1f9d98158c3f27f6ec6de45ee2874c9caf2a2c59 |
| Wan 2.1 I2V 14B 480P | Wan-AI/Wan2.1-I2V-14B-480P |
model | 6b73f84e66371cdfe870c72acd6826e1d61cf279 |
| FrameCrafter checkpoint | szqwu/FrameCrafter |
model | e5b2f6481a33a04ed7251d243784e16dda01996f |
| BSD 3ms/24ms archive | RuixuanJiang/Video_Deblurring_Datasets |
dataset | 62a6c6985c0c72caeaf372f821060ac442bdfa4e |
All six revisions were resolved successfully through the Hugging Face API on
2026-09-15 immediately before cleanup. source_metadata/ retains the local
Hugging Face download metadata, acquisition receipts, dataset manifests, and
preparation scripts needed to reconstruct the materialized layouts.
Important source hashes
| File | SHA-256 |
|---|---|
REDS train_blur.zip |
415c360c0d71d2d89af099b3c64e76bca7e1d8317750f8e056e02ebaab957bb8 |
REDS train_sharp.zip |
620294c1c3f23ed26c5ea228633770469c0b28e57d31eb41dc77deb401c6681b |
GoPro GOPRO_Large.zip |
24532c036712515cf33803704421311c91f6a080e76049d7e5e7fd22389f128e |
BSD BSD_3ms24ms.zip |
db14705307b4bcd75871c5efa832725a985dcdf398a63a737186ef0338a36ad2 |
FrameCrafter framecrafter.safetensors |
fa22ffdb9a54d5eb029618bcd424af1623ee6d64eb6277db7f7c478ab5b92524 |
Environment recovery
environment_specs/<environment>/ contains the explicit Conda package URL
list, full Conda YAML, history-only YAML, pip freeze, Python version, and
package JSON when available. Explicit files are the closest reconstruction
record; YAML and pip files make the record human-readable. Locally editable
packages still require their source checkout, which was not deleted by this
cleanup.
Audit files
planned_targets.json: pre-deletion path, allocation, category, and recovery basis for every selected target.deleted_targets.json: populated after deletion with the observed result.df_before.txtanddf_after.txt: filesystem-level proof of recovered space.MANIFEST.json: byte size and SHA-256 of every archive file except itself.
Never treat a cache or an installed environment directory as a scientific checkpoint. This archive preserves reconstruction metadata; unique learned weights and experiment outputs are deliberately not placed in the deletion set.
- Downloads last month
- 60