Dataset Viewer

Duplicate

Split (1)

test

The dataset viewer is not available for this split.

Cannot extract the features (columns) for the split 'test' of the config 'default' of the dataset.

Error code:   FeaturesError
Exception:    ArrowInvalid
Message:      JSON parse error: Column() changed from object to string in row 0
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 174, in _generate_tables
                  df = pandas_read_json(f)
                       ^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 38, in pandas_read_json
                  return pd.read_json(path_or_buf, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 815, in read_json
                  return json_reader.read()
                         ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1014, in read
                  obj = self._get_object_parser(self.data)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1040, in _get_object_parser
                  obj = FrameParser(json, **kwargs).parse()
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1176, in parse
                  self._parse()
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1391, in _parse
                  self.obj = DataFrame(
                             ^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/frame.py", line 778, in __init__
                  mgr = dict_to_mgr(data, index, columns, dtype=dtype, copy=copy, typ=manager)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/internals/construction.py", line 503, in dict_to_mgr
                  return arrays_to_mgr(arrays, columns, index, dtype=dtype, typ=typ, consolidate=copy)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/internals/construction.py", line 114, in arrays_to_mgr
                  index = _extract_index(arrays)
                          ^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/internals/construction.py", line 680, in _extract_index
                  raise ValueError(
              ValueError: Mixing dicts with non-Series may lead to ambiguous ordering.
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/split/first_rows.py", line 243, in compute_first_rows_from_streaming_response
                  iterable_dataset = iterable_dataset._resolve_features()
                                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 3496, in _resolve_features
                  features = _infer_features_from_batch(self.with_format(None)._head())
                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2257, in _head
                  return next(iter(self.iter(batch_size=n)))
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2461, in iter
                  for key, example in iterator:
                                      ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 1952, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 1974, in _iter_arrow
                  yield from self.ex_iterable._iter_arrow()
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 503, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 350, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 177, in _generate_tables
                  raise e
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 151, in _generate_tables
                  pa_table = paj.read_json(
                             ^^^^^^^^^^^^^^
                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
              pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to string in row 0

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.

YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/datasets-cards)

STT Fine-Tune Evaluation Results

Evaluation Date: December 10, 2025

This repository contains evaluation results comparing locally fine-tuned Whisper models against stock (original) Whisper inference across all model sizes.

Related Resources

Evaluation Script: Whisper-Fine-Tune-Accuracy-Eval
Evaluation Dataset: Small STT Eval Audio Dataset

Models Evaluated

Fine-tuned models: tiny, base, small, medium, large
Original (stock) models: tiny, base, small, medium, large

Results Summary

Model	Avg WER	Median WER	Min WER	Max WER	Avg Time/File (s)
original-medium	2.68%	0.00%	0.00%	14.29%	1.08
original-large	2.74%	0.00%	0.00%	15.79%	1.00
finetune-medium	2.93%	0.00%	0.00%	19.05%	1.47
finetune-large	2.97%	0.00%	0.00%	18.18%	0.99
finetune-small	3.41%	0.00%	0.00%	19.23%	0.82
original-small	3.92%	0.00%	0.00%	19.05%	0.52
finetune-base	4.97%	3.85%	0.00%	20.83%	0.41
original-base	6.49%	4.17%	0.00%	100.00%	0.41
finetune-tiny	6.55%	4.55%	0.00%	33.33%	0.25
original-tiny	7.76%	4.76%	0.00%	26.32%	0.28

Models sorted by average WER (best to worst)

Key Findings

Best overall performance: Original medium model (2.68% avg WER)
Fine-tuning benefit most visible at smaller model sizes: The fine-tuned base model significantly outperforms the original base model (4.97% vs 6.49% WER)
Original base model outlier: The original base model had a maximum WER of 100%, indicating complete transcription failures on some samples
92 audio files evaluated per model

Directory Structure

results/
├── summary_20251210_213315.csv          # Summary statistics for all models
├── detailed_results_20251210_213315.json # Full detailed results
├── evaluation_progress.json              # Evaluation progress tracking
└── individual/                           # Per-file results by model
    ├── finetune-base/
    ├── finetune-large/
    ├── finetune-medium/
    ├── finetune-small/
    ├── finetune-tiny/
    ├── original-base/
    ├── original-large/
    ├── original-medium/
    ├── original-small/
    └── original-tiny/

Metrics

WER (Word Error Rate): The primary metric used, calculated as (substitutions + insertions + deletions) / reference words
Duration: Inference time per audio file in seconds

Downloads last month: 1