README.md

---
license: other
task_categories:
  - audio-classification
pretty_name: SRE Dataset
configs:
  - config_name: default
    data_files:
      - split: test_sph
        path: "data/test_sph/audio/*.tar"
      - split: test_h
        path: "data/test_h/audio/*.tar"
      - split: test_other
        path: "data/test_other/audio/*.tar"
tags:
  - audio
  - speech
  - speaker-recognition
  - sre
  - webdataset
---

# Yougen/sre_testset

Speaker Recognition (SRE) speech dataset, packed as **WebDataset tar shards**.

## Layout

```
data/
  train/
    metadata.csv
    audio/
      train-000.tar
      train-001.tar
      ...
  validation/
    metadata.csv
    audio/
      validation-000.tar
      ...
  test/
    metadata.csv
    audio/
      test-000.tar
      ...
```

Shard counts:

- `test_sph`: 22 tar shard(s)
- `test_h`: 67 tar shard(s)
- `test_other`: 96 tar shard(s)

Inside each tar, every sample is a pair sharing a unique key:
```
<key>.wav       # raw audio bytes
<key>.json      # {"id":..., "rel_path":..., "wav_format":..., "duration":..., "label_str":..., "label":...}
```

`metadata.csv` columns:
`key, shard, id, rel_path, wav_format, duration, label_str, label`


## Loading

```python
from datasets import load_dataset

ds = load_dataset("Yougen/sre_testset")
print(ds)
print(ds["train"][0])
# sample keys: 'wav' (decoded audio), 'json' (metadata), '__key__', '__url__'
```

For streaming (no full download needed):

```python
ds = load_dataset("Yougen/sre_testset", streaming=True)
for example in ds["train"]:
    print(example["__key__"], example["json"]["label_str"])
    break
```

HuggingFace's `webdataset` builder will automatically pair `<key>.wav` with
`<key>.json` inside every tar and decode the audio.