Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    TypeError
Message:      int() argument must be a string, a bytes-like object or a real number, not 'NoneType'
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1539, in _prepare_split_single
                  for key, record 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/webdataset/webdataset.py", line 131, in _generate_examples
                  for example_idx, example in enumerate(self._get_pipeline_from_tar(tar_path, tar_iterator)):
                                              ~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 32, in _get_pipeline_from_tar
                  for filename, f in tar_iterator:
                                     ^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/track.py", line 49, in __iter__
                  for x in self.generator(*self.args):
                           ~~~~~~~~~~~~~~^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/file_utils.py", line 1419, in _iter_from_urlpath
                  with xopen(urlpath, "rb", download_config=download_config, block_size=0) as f:
                       ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/file_utils.py", line 976, in xopen
                  file_obj = fs.open(paths[0], mode)
                File "<string>", line 3, in open
                File "/usr/local/lib/python3.14/unittest/mock.py", line 1176, in __call__
                  return self._mock_call(*args, **kwargs)
                         ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/unittest/mock.py", line 1180, in _mock_call
                  return self._execute_mock_call(*args, **kwargs)
                         ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/unittest/mock.py", line 1247, in _execute_mock_call
                  result = effect(*args, **kwargs)
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 794, in wrapped
                  tracker.files[urlpath] = {"read": 0, "size": int(f.size)}
                                                               ~~~^^^^^^^^
              TypeError: int() argument must be a string, a bytes-like object or a real number, not 'NoneType'
              
              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 1401, 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 1579, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

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.

json
dict
__key__
string
__url__
string
{"video":"-DAp0zXkd6A","fps":0.5,"video_fps":24.0,"n":674,"variant":"pret_v2","repr":"pooled","promp(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/-DAp0zXkd6(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"-YO14FM76lY","fps":0.5,"video_fps":29.97003,"n":562,"variant":"pret_v2","repr":"pooled","p(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/-YO14FM76l(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"-l2_zQBeEOo","fps":0.5,"video_fps":60.0,"n":972,"variant":"pret_v2","repr":"pooled","promp(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/-l2_zQBeEO(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"-tpms_BbjHA","fps":0.5,"video_fps":23.976024,"n":614,"variant":"pret_v2","repr":"pooled","(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/-tpms_BbjH(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"-xlmBoQ4paI","fps":0.5,"video_fps":59.940063,"n":825,"variant":"pret_v2","repr":"pooled","(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/-xlmBoQ4pa(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"02LnK6fEdJw","fps":0.5,"video_fps":59.940063,"n":769,"variant":"pret_v2","repr":"pooled","(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/02LnK6fEdJ(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"07QhrrI5B4k","fps":0.5,"video_fps":29.97003,"n":732,"variant":"pret_v2","repr":"pooled","p(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/07QhrrI5B4(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"0JW8Ldb54qQ","fps":0.5,"video_fps":50.0,"n":894,"variant":"pret_v2","repr":"pooled","promp(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/0JW8Ldb54q(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"0P2dzIa6pZY","fps":0.5,"video_fps":30.0,"n":1221,"variant":"pret_v2","repr":"pooled","prom(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/0P2dzIa6pZ(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
{"video":"0__-AauBJz0","fps":0.5,"video_fps":50.0,"n":1063,"variant":"pret_v2","repr":"pooled","prom(...TRUNCATED)
"rekv_pret_captions_tsp_20261112/text_captions/cgbench/internvl3_5_8b/pooled/0.5fps__pret/0__-AauBJz(...TRUNCATED)
"hf://datasets/junipark/vatm-pret-captions@b449de82a84ce3cd61725daf8512e26cefe9d24c/rekv_pret_captio(...TRUNCATED)
End of preview.

VaTm P-RET captions

Per-frame (InternVL3.5-8B, LLaVA-OneVision-7B) and per-block (Qwen3-VL-8B) retrieval captions (P-RET prompt) used by the VaTm main runs on the juni_eval branch of the ReKV code. The captions were generated at 0.5 fps from the benchmark videos.

Files

File Size Contents
rekv_pret_captions_tsp_20261112.tar.xz 277 MB main bundle: 12 backbone × benchmark folders (see the table below)
rekv_pret_captions_tsp_20261112_qwen3_mlvu.tar.xz 17 MB add-on: Qwen3-VL-8B MLVU
rekv_pret_captions_tsp_20261112_lvbench_ivl_llava.tar.xz 23 MB add-on: LVBench for InternVL3.5-8B and LLaVA-OneVision-7B
*.sha256 checksums of the archives

Benchmarks and backbones

Which archive holds the captions for each benchmark and backbone.

Benchmark InternVL3.5-8B LLaVA-OneVision-7B Qwen3-VL-8B
LVBench _lvbench_ivl_llava add-on _lvbench_ivl_llava add-on main
CG-Bench main main main
MLVU (dev, without surveil_176) main main _qwen3_mlvu add-on
Video-MME medium main main main
Video-MME long main main main

All 15 benchmark × backbone pairs are covered once the three archives are unpacked.

Use

tar -xJf rekv_pret_captions_tsp_20261112.tar.xz -C /path/to
tar -xJf rekv_pret_captions_tsp_20261112_qwen3_mlvu.tar.xz -C /path/to          # same parent folder
tar -xJf rekv_pret_captions_tsp_20261112_lvbench_ivl_llava.tar.xz -C /path/to   # same parent folder
cd /path/to/rekv_pret_captions_tsp_20261112
sha256sum -c SHA256SUMS --quiet && sha256sum -c SHA256SUMS_qwen3_mlvu --quiet && sha256sum -c SHA256SUMS_lvbench_ivl_llava --quiet
export REKV_CAPTION_ROOT=/path/to/rekv_pret_captions_tsp_20261112/text_captions

Layout: text_captions/<benchmark>/<captioner>/pooled/<variant>/<video_id>.json. MANIFEST.tsv (and MANIFEST_lvbench_ivl_llava.tsv for the LVBench add-on) lists which run set reads which folder. The [mm:ss] timestamp prefix is added by the code at run time; it is not stored in the files.

InternVL3.5-8B LVBench reads two folders: 0.5fps__pret_raw128 (captions generated with a 128-token limit) and _pret_recap/lvbench (the frames cut at 128 tokens, generated again with a 256-token limit). The code merges them into the 256-token caption set.

License

The captions are derived from the videos of the benchmarks above. Use them for research only and follow the license and terms of each benchmark.

Downloads last month
97