The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
Error code: DatasetGenerationError
Exception: UnicodeDecodeError
Message: 'utf-8' codec can't decode byte 0x80 in position 64: invalid start byte
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1835, in _prepare_split_single
for key, table in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/text/text.py", line 98, in _generate_tables
batch = f.read(self.config.chunksize)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/file_utils.py", line 838, in read_with_retries
out = read(*args, **kwargs)
File "<frozen codecs>", line 325, in decode
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 64: invalid start byte
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/parquet_and_info.py", line 1379, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 956, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1702, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1887, in _prepare_split_single
raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the datasetNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
text string |
|---|
"bpe_simple_vocab_16e6.txt#version: 0.2 |
i n |
t h |
a n |
r e |
a r |
e r |
th e</w> |
in g</w> |
o u |
o n |
s t |
o r |
e n |
o n</w> |
a l |
a t |
e r</w> |
i t |
i n</w> |
t o</w> |
r o |
i s</w> |
l e |
i c |
a t</w> |
an d</w> |
e d</w> |
o f</w> |
c h |
o r</w> |
e s</w> |
i l |
e l |
s t</w> |
a c |
o m |
a m |
l o |
a n</w> |
a y</w> |
s h |
r i |
l i |
t i |
f or</w> |
n e |
ð Ł |
r a |
h a |
d e |
o l |
v e</w> |
s i |
u r |
a l</w> |
s e |
' s</w> |
u n |
d i |
b e |
l a |
w h |
o o |
d ay</w> |
e n</w> |
m a |
n o |
l e</w> |
t o |
ou r</w> |
i r |
g h |
w it |
i t</w> |
y o |
a s |
s p |
th is</w> |
t s</w> |
at i |
yo u</w> |
wit h</w> |
a d |
i s |
a b |
l y</w> |
w e |
th e |
t e |
a s</w> |
a g |
v i |
p p |
s u |
h o |
m y</w> |
. . |
b u |
c om |
VBench evaluation auxiliary assets
Public mirror of 18 auxiliary files used by ElasticTok's pinned VBench and VBench-2.0 T2V evaluators: 3,844,791,432 bytes (3.84 GB). These are original upstream files, byte-for-byte verified against the evaluator's existing SHA256 pins. This repository changes download locations, not model weights or scoring rules.
The mirror enables evaluation preparation on clusters that can access Hugging Face and GitHub but cannot access Google Drive, Dropbox, Meta's download host, Azure's CLIP host, or the PyTorch model download host.
Hanlin / ElasticTok usage
Update the ElasticTok checkout and follow the complete T2V guide. After installing dependencies, run on a CPU/login node:
python eval_vbench_t2v.py --suite vbench --stage prepare --base-dir "$BASE"
python eval_vbench_t2v.py --suite vbench2 --stage prepare --base-dir "$BASE"
The updated preparer downloads the files in this mirror at a pinned Hub commit and validates every SHA256. Models already available on HF (including Qwen, LLaVA, SigLIP, BERT and other scoring weights) continue to use their upstream repositories. Pinned source checkouts still come from GitHub; package installation uses the normal environment setup. All generation/scoring runs use local assets.
This is not the complete evaluator environment or the complete model cache. VBench's complete scoring cache is approximately 9 GB; VBench2's is approximately 42 GB. Generation-model caches are separate. ElasticTok training checkpoints and generated benchmark videos are not included.
Contents and verification
assets/: cache-relative paths for DINO, CLIP, aesthetic prediction, MUSIQ, ViCLIP vocabulary, RAFT, VGG19, RetinaFace, CoTracker2, and VBench2 anomaly/ identity detector assets.manifest.json: each file's original source URL, byte length, SHA256 and suites.checksums.sha256: verify a downloaded snapshot withsha256sum -c checksums.sha256.ATTRIBUTION.mdandlicenses/: source attribution and available upstream notices.
RAFT is the original extracted raft-things.pth member of the upstream archive.
The RetinaFace ZIP is mirrored unchanged. The instance anomaly detector adapter
includes all four pinned files, including its original configuration metadata.
To download the mirror itself without running the integrated preparer:
hf download xcpan/vbench_eval --repo-type dataset --local-dir ./vbench_eval_mirror
cd vbench_eval_mirror
sha256sum -c checksums.sha256
Attribution and terms
This collection does not grant a new blanket license over upstream weights. Original authors retain their rights and source-specific terms apply, including noncommercial restrictions where specified. See ATTRIBUTION.md and the linked upstream projects before using or redistributing individual files. This is a convenience mirror for evaluation, not an official VBench release.
Upstream evaluator: Vchitect/VBench, pinned
at fd18b3d055cb0fc6f066ca90fe2c3c8cbb698490 for this asset collection.
- Downloads last month
- 16