Datasets:

suryadv
/

dynsuperclevr-24-frame-resize256-eval-viewer

Error code:   StreamingRowsError
Exception:    ValueError
Message:      Invalid string class label dynsuperclevr-24-frame-resize256-eval-viewer@2da6966d193b547f9aae951ec7eedbbf6c4215fc
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 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2386, in __iter__
                  example = _apply_feature_types_on_example(
                      example, self.features, token_per_repo_id=self.token_per_repo_id
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2303, in _apply_feature_types_on_example
                  encoded_example = features.encode_example(example)
                File "/usr/local/lib/python3.14/site-packages/datasets/features/features.py", line 2178, in encode_example
                  return encode_nested_example(self, example)
                File "/usr/local/lib/python3.14/site-packages/datasets/features/features.py", line 1460, in encode_nested_example
                  {k: encode_nested_example(schema[k], obj.get(k), level=level + 1) for k in schema}
                      ~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/features/features.py", line 1483, in encode_nested_example
                  return schema.encode_example(obj) if obj is not None else None
                         ~~~~~~~~~~~~~~~~~~~~~^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/features/features.py", line 1158, in encode_example
                  example_data = self.str2int(example_data)
                File "/usr/local/lib/python3.14/site-packages/datasets/features/features.py", line 1095, in str2int
                  output = [self._strval2int(value) for value in values]
                            ~~~~~~~~~~~~~~~~^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/features/features.py", line 1116, in _strval2int
                  raise ValueError(f"Invalid string class label {value}")
              ValueError: Invalid string class label dynsuperclevr-24-frame-resize256-eval-viewer@2da6966d193b547f9aae951ec7eedbbf6c4215fc

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.

DynSuperCLEVR 24-frame evaluation viewer bundle

Open the hosted viewer · Download the complete ZIP

The ZIP contains the reusable static viewer and 40 selected examples. Run: Without mixture + mixture checkpoint-414 · 256×256 · PX-only · 20260912T015030Z. For each model independently: 10 cases starting at the 25th percentile of non-zero interaction F1s, plus the 10 highest-scoring cases, pooling the actual DynSuperCLEVR validation/test splits (99 + 96 per model). Use the model selector to switch between without-mixture and mixture results when the bundle contains both. Full semantic GT names (e.g. yellow mountain bicycle) and the agreed standalone aliases come from the corrected labels. For the resized run, RGB comes from the actual 256x256 inference videos; GT masks are resized only for display, and normalized PX is unchanged. This is a selected qualitative audit, not the full benchmark or a training set.

It includes 24 RGB and segmentation frames per case, exact raw predictions, reference labels, saved evaluator matches, per-sample scores, and source hashes. The score is the mean of temporal/input/output F1 at 0.25/0.5/0.75/1.0. manifest.json records the exact selection and caption-provenance hash. The visualization build does not rerun inference or evaluation. These assets, including raw predictions and reference labels, are public.

Run locally

Download visualizer.zip and unzip it.
Open a terminal inside the extracted dyn24-visualizer folder.
Run python3 serve.py (Windows: py serve.py). No pip installation needed.
Open http://127.0.0.1:8127/ in your browser.

The ZIP includes instructions for adding other examples. Its internal checksums.sha256 verifies the extracted files; visualizer.zip.sha256 in this dataset verifies the download itself.

GT crosses, predicted circles, object-specific toggles, and saved caption correspondence are synchronized across both images. Occlusion handling is a display rule, not an alteration of the raw predictions or saved evaluation.

Downloads last month: 36

Total file size:

105 MB