Add speaker-disjoint ASR re-split (v2): ach

## What this is

A speaker-disjoint re-split of the **labelled ASR splits** of this dataset,
added under `data/ASR_v2/` as new `{lang}_asr_v2` configs.

**Nothing existing is modified.** No file outside `data/ASR_v2/` is changed
except the root `README.md`, and that edit only *appends* the new config
entries -- every one of the existing configs and the entire prose body are
byte-identical, verified programmatically before this commit was built. Anyone
currently loading `{lang}_asr` (hackathons included) is unaffected.

The audio payloads are copied byte-for-byte from the current revision. This is
a re-partitioning, not a re-encode and not a quality filter.

## Why

The v1 ASR splits share speakers between train and eval, in several languages
almost completely:

| Language | v1 eval speakers also in train | v2 |
|---|---|---|
| Acholi (`ach`) | 99 % | **0 %** |
| Akan (`aka`) | 100 % | **0 %** |
| Dagbani (`dag`) | 98 % | **0 %** |
| Dagaare (`dga`) | 100 % | **0 %** |
| Ewe (`ewe`) | 100 % | **0 %** |
| Fula (`ful`) | 99 % | **0 %** |
| Ikposo (`kpo`) | 99 % | **0 %** |
| Lingala (`lin`) | 99 % | **0 %** |
| Luganda (`lug`) | 100 % | **0 %** |
| Masaaba (`mas`) | 98 % | **0 %** |
| Malagasy (`mlg`) | 98 % | **0 %** |
| Runyankole (`nyn`) | 98 % | **0 %** |
| Shona (`sna`) | 97 % | **0 %** |
| Soga (`sog`) | 98 % | **0 %** |

A model that has heard the eval speakers in training reports a score that is
partly speaker memorisation, and the size of that effect cannot be recovered
after the fact.

## Method

Speakers are the unit of assignment, so disjointness is structural rather than
enforced. Subject to that, the assignment minimises train/eval *rare-content*
overlap (idf-weighted, so shared common vocabulary -- which is coverage, not
leakage -- is deliberately invisible to the objective) and matches the gender
distribution of each eval split to the corpus. Deterministic from one seed;
full provenance in `data/ASR_v2/metadata/splits_manifest.json`.

## Storage note

Hugging Face PRs push to `refs/pr/N` on this repository, so these LFS objects
occupy your storage from now, before you have agreed to take them. If you would
rather not host them, closing the PR and deleting the ref reclaims the space
immediately -- and `data/ASR_v2/metadata/split_map.parquet` (a few MB) is
enough for anyone to reproduce the split locally without the audio at all.

Happy to adjust anything, or to reduce this to the metadata-only form.

data/ASR_v2/README.md

@@ -0,0 +1,292 @@
+---
+language:
+- ach
+- aka
+- amh
+- dag
+- dga
+- ewe
+- ful
+- kpo
+- lin
+- lug
+- mlg
+- myx
+- nyn
+- orm
+- sid
+- sna
+- tir
+- wal
+- xog
+license:
+- CC-BY-4.0
+- CC-BY-SA-4.0
+task_categories:
+- automatic-speech-recognition
+source_datasets:
+- google/WaxalNLP
+multilinguality:
+- multilingual
+annotation_creators:
+- human-annotated
+size_categories:
+- 100K<n<1M
+pretty_name: WAXAL ASR — speaker-disjoint re-split (v2)
+tags:
+- audio
+- automatic-speech-recognition
+- african-languages
+- speaker-disjoint-splits
+- resplit
+---
+
+# WAXAL ASR — speaker-disjoint re-split (v2)
+
+An **unofficial, community-contributed** re-split of the ASR half of
+[`google/WaxalNLP`](https://huggingface.co/datasets/google/WaxalNLP).
+Not affiliated with, endorsed by, or produced by Google.
+
+**The audio is unchanged.** Every payload is the byte-for-byte MP3 from v1;
+this is a re-partitioning of the same recordings, not a re-encode and not a
+quality filter. What changes is *which utterances are in which split*.
+
+- 438,230 labelled utterances, 2,242 hours,
+  19 languages
+- v1 remains untouched and fully usable: `load_dataset("google/WaxalNLP", "ach_asr")`
+- v2 is a separate config: `load_dataset("google/WaxalNLP", "ach_asr_v2")`
+
+## Why
+
+The v1 ASR splits are not safe to benchmark on. Measured across all
+19 languages:
+
+| Code | Language | v1 eval speakers also in train | v2 | v1 eval transcripts also in train | v2 | v2 leakage vs chance |
+|---|---|---|---|---|---|---|
+| `ach` | Acholi | 99.0 % | **0.0 %** | 0.6 % | 1.5 % | 0.88 (-2.8σ) |
+| `aka` | Akan | 100.0 % | **0.0 %** | 0.0 % | 0.0 % | 0.95 (-0.9σ) |
+| `amh` | Amharic | 0.0 % | **0.0 %** | 0.1 % | 0.4 % | 0.70 (-6.0σ) |
+| `dag` | Dagbani | 97.9 % | **0.0 %** | 0.2 % | 0.1 % | 0.75 (-9.3σ) |
+| `dga` | Dagaare | 100.0 % | **0.0 %** | 19.1 % | 0.2 % | 0.83 (-3.0σ) |
+| `ewe` | Ewe | 99.8 % | **0.0 %** | 0.0 % | 0.0 % | 0.86 (-3.5σ) |
+| `ful` | Fula | 99.4 % | **0.0 %** | 0.0 % | 0.0 % | 0.97 (-0.5σ) |
+| `kpo` | Ikposo | 99.1 % | **0.0 %** | 0.0 % | 0.1 % | 0.76 (-4.4σ) |
+| `lin` | Lingala | 98.7 % | **0.0 %** | 0.1 % | 0.1 % | 1.06 (+1.0σ) |
+| `lug` | Luganda | 99.6 % | **0.0 %** | 0.3 % | 0.3 % | 0.82 (-5.8σ) |
+| `mas` | Masaaba | 98.4 % | **0.0 %** | 0.2 % | 0.2 % | 0.80 (-4.8σ) |
+| `mlg` | Malagasy | 97.7 % | **0.0 %** | 0.0 % | 0.0 % | 0.90 (-1.4σ) |
+| `nyn` | Runyankole | 98.5 % | **0.0 %** | 0.7 % | 0.5 % | 0.78 (-6.1σ) |
+| `orm` | Oromo | 0.0 % | **0.0 %** | 0.0 % | 0.2 % | 0.51 (-7.9σ) |
+| `sid` | Sidama | 0.0 % | **0.0 %** | 0.0 % | 0.2 % | 0.67 (-4.8σ) |
+| `sna` | Shona | 97.0 % | **0.0 %** | 0.1 % | 0.1 % | 1.09 (+1.0σ) |
+| `sog` | Soga | 98.3 % | **0.0 %** | 0.0 % | 0.0 % | 0.87 (-3.8σ) |
+| `tir` | Tigrinya | 0.0 % | **0.0 %** | 2.0 % | 0.1 % | 0.69 (-4.7σ) |
+| `wal` | Wolaytta | 0.0 % | **0.0 %** | 0.0 % | 0.1 % | 0.83 (-3.5σ) |
+
+A model that has heard the test speakers during training reports a score that
+is partly speaker memorisation rather than recognition, and the size of that
+effect is not knowable after the fact. The re-split removes the channel rather
+than trying to correct for it.
+
+## How the splits were built
+
+Speakers are the unit of assignment, so **speaker-disjointness is structural**:
+an utterance cannot cross a split boundary without its speaker, and a speaker
+belongs to exactly one split. The assignment is then chosen to satisfy, in
+order of precedence:
+
+1. **Speaker-disjointness** — a hard constraint, enforced by construction.
+2. **Low train/eval lexical leakage** — a `sqrt(idf)`-weighted affinity over
+   shared sentences, word trigrams and word types is minimised across the
+   train/eval boundary. Because inverse document frequency goes to zero for
+   items nearly everyone uses, *shared common vocabulary is invisible to the
+   objective* and only shared rare content is penalised. This is deliberate:
+   common-word overlap between train and test is coverage, not leakage, and a
+   test set that avoided it would measure a distribution nobody deploys against.
+3. **Gender balance** — the female share of each eval split is matched to the
+   corpus, weighted by reference words (the weighting under which a
+   micro-averaged WER is or is not gender-representative).
+
+Targets are 80/10/10 by duration, but two bounds move that in practice, and
+both are recorded per language in `splits_manifest.json`: the eval share is
+allowed to **grow** to reach a 20,000-reference-word floor in the smallest
+languages, and is **capped** at 12 hours or 2,000 utterances per eval split in
+the largest — so a 220-hour language like Amharic evaluates on about 5.5 % per
+side rather than 10 %, with the surplus going to train. No single speaker may
+exceed 25 % of an eval split. The search is a randomised
+longest-processing-time greedy seed followed by steepest-descent local search,
+seeded from `20260730` and fully deterministic; re-running reproduces
+the assignment hash recorded per language in `splits_manifest.json`.
+
+Every goal is scaled against a **null distribution** of
+200 random size-matched speaker-disjoint splits
+of the same language, so the reported numbers can be read against chance.
+
+## Splits
+
+Train / validation / test, as utterances, hours and distinct speakers:
+
+| Code | train | validation | test |
+|---|---|---|---|
+| `ach` | 4,120 / 26.0 h / 263 | 520 / 3.2 h / 33 | 515 / 3.2 h / 33 |
+| `aka` | 10,196 / 55.6 h / 110 | 1,287 / 7.0 h / 11 | 1,269 / 6.9 h / 13 |
+| `amh` | 39,455 / 196.0 h / 550 | 2,473 / 12.0 h / 34 | 2,426 / 12.0 h / 34 |
+| `dag` | 14,237 / 77.2 h / 884 | 1,798 / 9.7 h / 79 | 1,784 / 9.7 h / 107 |
+| `dga` | 15,091 / 83.8 h / 275 | 1,887 / 10.5 h / 36 | 1,896 / 10.5 h / 38 |
+| `ewe` | 15,092 / 79.8 h / 431 | 1,885 / 10.0 h / 53 | 1,884 / 10.0 h / 54 |
+| `ful` | 19,293 / 100.6 h / 168 | 2,303 / 12.0 h / 21 | 2,285 / 11.9 h / 22 |
+| `kpo` | 14,416 / 82.3 h / 369 | 1,803 / 10.3 h / 46 | 1,800 / 10.3 h / 46 |
+| `lin` | 14,523 / 72.3 h / 74 | 1,778 / 9.1 h / 13 | 1,808 / 9.1 h / 10 |
+| `lug` | 5,366 / 36.7 h / 270 | 692 / 4.7 h / 35 | 699 / 4.7 h / 35 |
+| `mas` | 6,856 / 39.2 h / 290 | 857 / 4.9 h / 36 | 861 / 4.9 h / 36 |
+| `mlg` | 18,467 / 94.5 h / 199 | 2,316 / 11.8 h / 25 | 2,303 / 11.8 h / 27 |
+| `nyn` | 6,752 / 40.8 h / 293 | 861 / 5.1 h / 37 | 859 / 5.2 h / 37 |
+| `orm` | 40,141 / 198.7 h / 361 | 2,473 / 11.9 h / 23 | 2,431 / 12.0 h / 22 |
+| `sid` | 40,547 / 202.0 h / 463 | 2,444 / 11.9 h / 28 | 2,413 / 12.0 h / 28 |
+| `sna` | 14,079 / 79.5 h / 131 | 1,744 / 9.9 h / 20 | 1,762 / 9.9 h / 17 |
+| `sog` | 6,298 / 40.4 h / 260 | 786 / 5.1 h / 33 | 793 / 5.0 h / 32 |
+| `tir` | 44,945 / 194.8 h / 408 | 2,686 / 12.1 h / 25 | 2,684 / 12.1 h / 24 |
+| `wal` | 42,258 / 195.1 h / 381 | 2,572 / 12.2 h / 23 | 2,461 / 12.0 h / 23 |
+
+### Gender balance
+
+Female share of reference words. `n/a` means the source has **no gender labels
+at all** for that language — not that balance was achieved.
+
+| Code | train | validation | test |
+|---|---|---|---|
+| `ach` | 0.268 | 0.267 | 0.264 |
+| `aka` | n/a | n/a | n/a |
+| `amh` | 0.454 | 0.454 | 0.454 |
+| `dag` | n/a | n/a | n/a |
+| `dga` | n/a | n/a | n/a |
+| `ewe` | n/a | n/a | n/a |
+| `ful` | n/a | n/a | n/a |
+| `kpo` | n/a | n/a | n/a |
+| `lin` | 0.550 | 0.548 | 0.539 |
+| `lug` | 0.470 | 0.471 | 0.470 |
+| `mas` | 0.306 | 0.307 | 0.306 |
+| `mlg` | n/a | n/a | n/a |
+| `nyn` | 0.452 | 0.454 | 0.452 |
+| `orm` | 0.562 | 0.561 | 0.559 |
+| `sid` | 0.654 | 0.656 | 0.660 |
+| `sna` | 0.614 | 0.612 | 0.600 |
+| `sog` | 0.479 | 0.477 | 0.482 |
+| `tir` | 0.552 | 0.556 | 0.552 |
+| `wal` | 0.661 | 0.661 | 0.656 |
+
+### Languages needing attention
+
+| Code | Status | Note |
+|---|---|---|
+| `ach` | ok | test: 19,399 reference words against the 20,000 target (WER standard error ~0.36 pp before clustering) |
+| `amh` | ok | eval share reduced 0.100 -> 0.055 by the 12-hour eval cap |
+| `ful` | ok | eval share reduced 0.100 -> 0.096 by the 12-hour eval cap |
+| `lug` | ok | eval share raised 0.100 -> 0.102 to reach the 20,000-reference-word floor; validation: 18,292 reference words against the 20,000 target (WER standard error ~0.37 pp before clustering); test: 18,928 reference words against the 20,000 target (WER standard error ~0.36 pp before clustering) |
+| `mas` | ok | validation: 19,731 reference words against the 20,000 target (WER standard error ~0.35 pp before clustering); test: 14,265 reference words against the 20,000 target (WER standard error ~0.41 pp before clustering) |
+| `nyn` | ok | eval share raised 0.100 -> 0.101 to reach the 20,000-reference-word floor; validation: 17,439 reference words against the 20,000 target (WER standard error ~0.37 pp before clustering); test: 19,041 reference words against the 20,000 target (WER standard error ~0.36 pp before clustering) |
+| `orm` | ok | eval share reduced 0.100 -> 0.054 by the 12-hour eval cap |
+| `sid` | ok | eval share reduced 0.100 -> 0.053 by the 12-hour eval cap |
+| `tir` | ok | eval share reduced 0.100 -> 0.055 by the 12-hour eval cap |
+| `wal` | ok | eval share reduced 0.100 -> 0.055 by the 12-hour eval cap |
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("google/WaxalNLP", "sna_asr_v2")
+print(ds["test"][0]["transcription"])
+```
+
+To re-split a copy of v1 you already have, without downloading anything else,
+join on `(language, id)` against `data/ASR_v2/metadata/split_map.parquet`:
+
+```python
+import pandas as pd
+mapping = pd.read_parquet("split_map.parquet")          # language, id, v1_split, v2_split
+v2_split = mapping.set_index(["language", "id"])["v2_split"]
+```
+
+## Schema
+
+The six v1 columns, unchanged, plus derived fields:
+
+| Column | Notes |
+|---|---|
+| `id`, `speaker_id`, `transcription`, `language`, `gender` | unchanged from v1 |
+| `audio` | `struct<bytes, path>` — **byte-identical** to v1 |
+| `v1_split` | which v1 split this utterance came from |
+| `duration_s`, `n_bytes` | derived; the corpus is 128 kbit/s CBR MP3 |
+| `sample_rate`, `channels` | measured per shard; **heterogeneous** (see limitations) |
+| `provider`, `licence` | the licence travels with the row |
+
+The stray `__index_level_0__` column present in 36 of the v1 shards is dropped,
+and every v2 file uses 100-row row groups.
+
+## Limitations
+
+- **No gender labels at all** for `aka`, `dag`, `dga`, `ewe`, `ful`, `kpo`, `mlg`.
+  The gender goal is undefined there, not satisfied, and is reported as `n/a`.
+- **Heterogeneous audio format.** The corpus mixes 16 kHz mono and 48 kHz
+  stereo MP3 between languages. v1 documents neither; v2 records the measured
+  values per row but does **not** resample, because a silent normalisation
+  cannot be verified against the original.
+- **This is a re-split, not a clean-up.** No quality filtering has been
+  applied. Utterance-level defects present in v1 are present here.
+- **The `unlabeled` split is out of scope** (~816 GB, no transcripts). Take it
+  from the v1 config, which is unchanged.
+- **`speaker_id` is trusted, not verified acoustically** unless the manifest
+  says otherwise. If an id turns out to cover more than one voice, the
+  disjointness guarantee is over ids, not over voices.
+- The `v1 eval transcripts also in train` column is an **exact census** over
+  every eval transcript. A separate fuzzy near-duplicate rate (0.95 similarity)
+  is recorded in `metadata/split_comparison.csv` and is computed on a sample of
+  at most 1,200 eval transcripts per split, with the sample
+  size stored beside it.
+- Rows with an empty or placeholder transcript, or with no `speaker_id`, are
+  excluded from the splits and listed in `metadata/excluded_rows.parquet`.
+  Nothing else is filtered.
+
+## Licensing
+
+The corpus mixes two licences by contributing partner. **Each language is a
+separate config specifically so that aggregating them does not relicense the
+CC-BY-4.0 languages under ShareAlike.** The `licence` column travels with every
+row.
+
+| Provider | Languages | Licence |
+|---|---|---|
+| Digital Umuganda | Amharic (`amh`), Fula (`ful`), Lingala (`lin`), Malagasy (`mlg`), Oromo (`orm`), Shona (`sna`), Sidama (`sid`), Tigrinya (`tir`), Wolaytta (`wal`) | `CC-BY-SA-4.0` |
+| Makerere University | Acholi (`ach`), Luganda (`lug`), Masaaba (`mas`), Runyankole (`nyn`), Soga (`sog`) | `CC-BY-SA-4.0` |
+| University of Ghana | Akan (`aka`), Dagaare (`dga`), Dagbani (`dag`), Ewe (`ewe`), Ikposo (`kpo`) | `CC-BY-4.0` |
+
+Derivatives of the CC-BY-SA-4.0 languages must carry the same terms. Combining
+configs means inheriting ShareAlike for the combined work.
+
+## Attribution and citation
+
+Source corpus: [`google/WaxalNLP`](https://huggingface.co/datasets/google/WaxalNLP),
+revision `e0a62aaebc61`, collected by Makerere
+University, the University of Ghana and Digital Umuganda, funded by Google and
+the Gates Foundation. Upstream sources:
+[UGSpeechData](https://doi.org/10.57760/sciencedb.22298),
+[AfriVoice](https://huggingface.co/datasets/DigitalUmuganda/AfriVoice), and the
+[Yogera Dataset](https://doi.org/10.7910/DVN/BEROE0).
+
+```bibtex
+@article{waxal2026,
+  title={WAXAL: A Large-Scale Multilingual African Language Speech Corpus},
+  author={Anonymous},
+  journal={arXiv preprint arXiv:2602.02734},
+  year={2026}
+}
+```
+
+## Reproducing this
+
+The split is fully determined by the manifest. `splits_manifest.json` records
+the seed, the pinned source revision, every tolerance, the library versions,
+and a hash of the text-canonicalisation function that all leakage figures were
+computed through. `splits_manifest.parquet` carries the per-utterance
+assignment, and `shard_manifest.parquet` the SHA-256 of every written shard.

data/ASR_v2/metadata/split_comparison.csv

@@ -0,0 +1,115 @@
+language,version,split,n_utterances,hours,n_speakers,ref_words,female_share,female_share_words,unknown_gender_share,speaker_leak_from_train,exact_sentence_overlap,near_dup_overlap,near_dup_sample,oov_token_rate,oov_type_rate,unseen_char_rate,max_speaker_share
+ach,v1,train,4108,25.848990885416665,322,160803,0.30331061343719573,0.2632227010690099,0.0,,,,,,,,0.01802406832575798
+ach,v1,validation,519,3.310818142361111,199,20849,0.3140655105973025,0.2894623243321023,0.0,0.9748743718592965,0.0038535645472061657,0.0038535645472061657,519.0,0.030169312676866997,0.22230858919549165,0.0,0.019997138530015945
+ach,v1,test,528,3.2940662977430555,194,20627,0.3181818181818182,0.28196053716003294,0.0,0.9896907216494846,0.005681818181818182,0.005681818181818182,528.0,0.02860328695399234,0.21675845790715972,0.037037037037037035,0.026069996878504753
+ach,v2,train,4120,25.965470920138888,263,162649,0.3048543689320388,0.26844308910598896,0.0,,,,,,,,0.022207921370863914
+ach,v2,validation,520,3.2417279730902777,33,20231,0.26346153846153847,0.2669171074094212,0.0,0.0,0.015384615384615385,0.015384615384615385,520.0,0.03000346003657753,0.22743977133523888,0.0,0.13066771626472473
+ach,v2,test,515,3.2466758897569443,33,19399,0.3572815533980582,0.26372493427496263,0.0,0.0,0.0,0.0,515.0,0.029743801226867365,0.21784666945957268,0.0,0.10795231908559799
+aka,v1,train,10107,55.02978298611111,101,348350,,,1.0,,,,,,,,0.04007843881845474
+aka,v1,validation,1123,6.153727213541667,99,39344,,,1.0,1.0,0.0,0.009795191451469279,1123.0,0.02406974379829199,0.21604330708661418,0.0,0.04804767668247223
+aka,v1,test,1522,8.363488498263889,33,51740,,,1.0,0.0,0.0,0.0033333333333333335,1200.0,0.029010436799381523,0.2822512647554806,0.030303030303030304,0.033697500824928284
+aka,v2,train,10196,55.60658854166667,110,348373,,,1.0,,,,,,,,0.04497991502285004
+aka,v2,validation,1287,7.02416015625,11,48085,,,1.0,0.0,0.0,0.0025,1200.0,0.027347405635853177,0.2653995878177238,0.0,0.14713680744171143
+aka,v2,test,1269,6.91625,13,42976,,,1.0,0.0,0.0,0.015,1200.0,0.04099962769918094,0.3070263740688544,0.058823529411764705,0.11151575297117233
+amh,v1,train,38022,189.7320138888889,566,1007363,0.4307769186260586,0.43713040880000553,0.0,,,,,,,,0.009729607030749321
+amh,v1,validation,2912,13.968602430555556,28,77783,0.5233516483516484,0.5303343918336911,0.0,0.0,0.0010302197802197802,0.0016666666666666668,1200.0,0.05361068613964491,0.2319344571495933,0.0,0.1053454726934433
+amh,v1,test,3420,16.2302734375,24,91866,0.5701754385964912,0.578973722595955,0.0,0.0,0.0011695906432748538,0.008333333333333333,1200.0,0.0595976748742734,0.2655541069100391,0.0,0.0884522795677185
+amh,v2,train,39455,195.99618055555555,550,1060099,0.44795336459257384,0.4544169931298869,0.0,,,,,,,,0.009418642148375511
+amh,v2,validation,2473,11.968304036458333,34,57381,0.47594015365952286,0.4540353078545163,0.0,0.0,0.003234937323089365,0.0033333333333333335,1200.0,0.07385720011850612,0.2842906968467507,0.007575757575757576,0.09336347132921219
+amh,v2,test,2426,11.966408420138889,34,59532,0.41302555647155814,0.4536719747362763,0.0,0.0,0.003709810387469085,0.0025,1200.0,0.06720755224081167,0.24459538893605903,0.0,0.12144400179386139
+dag,v1,train,14231,77.17661458333333,1046,589860,,,1.0,,,,,,,,0.01278792042285204
+dag,v1,validation,1750,9.509487847222223,642,73607,,,1.0,0.9766355140186916,0.0017142857142857142,0.005,1200.0,0.01695490917983344,0.2249464876435104,0.0,0.015343086794018745
+dag,v1,test,1838,9.912782118055556,682,75234,,,1.0,0.9794721407624634,0.000544069640914037,0.0033333333333333335,1200.0,0.017385756439907488,0.22911051212938005,0.023809523809523808,0.00924919918179512
+dag,v2,train,14237,77.18172743055555,884,597247,,,1.0,,,,,,,,0.007015130016952753
+dag,v2,validation,1798,9.744738498263889,79,69418,,,1.0,0.0,0.0005561735261401557,0.0016666666666666668,1200.0,0.017300988216312773,0.23394801155298822,0.05,0.12565946578979492
+dag,v2,test,1784,9.672421875,107,72036,,,1.0,0.0,0.0005605381165919282,0.0025,1200.0,0.025237381309345327,0.2863587152652472,0.05,0.05325062945485115
+dga,v1,train,15071,83.60558159722223,348,584154,,,1.0,,,,,,,,0.03391222283244133
+dga,v1,validation,1893,10.551019965277778,264,74584,,,1.0,0.9962121212121212,0.19070258848388802,0.19333333333333333,1200.0,0.021680253137402124,0.22939690558888537,0.0,0.04143372178077698
+dga,v1,test,1910,10.525425347222223,261,73027,,,1.0,1.0,0.16020942408376965,0.1775,1200.0,0.023922658742656824,0.2523483521732208,0.014925373134328358,0.03608813136816025
+dga,v2,train,15091,83.75778645833333,275,593106,,,1.0,,,,,,,,0.042447760701179504
+dga,v2,validation,1887,10.451787109375,36,68704,,,1.0,0.0,0.0010598834128245894,0.006666666666666667,1200.0,0.04376746623195156,0.36274818783485663,0.014925373134328358,0.1290062516927719
+dga,v2,test,1896,10.472458767361111,38,69955,,,1.0,0.0,0.002109704641350211,0.0075,1200.0,0.03989707669215924,0.3380831212892282,0.030303030303030304,0.10429682582616806
+ewe,v1,train,15054,79.67143229166666,533,498586,,,1.0,,,,,,,,0.028677133843302727
+ewe,v1,validation,1916,10.11669162326389,423,63074,,,1.0,0.9905437352245863,0.0,0.0008333333333333334,1200.0,0.027761042584900277,0.27922624053826745,0.0,0.0249943770468235
+ewe,v1,test,1891,10.009307725694445,401,62919,,,1.0,0.9975062344139651,0.0,0.0008333333333333334,1200.0,0.028496956404265802,0.28683092608326255,0.0,0.025948595255613327
+ewe,v2,train,15092,79.84296006944444,431,498576,,,1.0,,,,,,,,0.035035476088523865
+ewe,v2,validation,1885,9.979256727430556,53,62132,,,1.0,0.0,0.0,0.0,1200.0,0.031239940771261184,0.31020196789228377,0.019230769230769232,0.1145356222987175
+ewe,v2,test,1884,9.975211588541667,54,63871,,,1.0,0.0,0.0,0.0,1200.0,0.030248469571480016,0.30514266188279515,0.0,0.06748399883508682
+ful,v1,train,19132,99.73537326388889,208,627206,,,1.0,,,,,,,,0.046491872519254684
+ful,v1,validation,2358,12.255303819444444,163,77195,,,1.0,0.9877300613496932,0.0,0.0,1200.0,0.047490122417255004,0.3185198958720714,0.0392156862745098,0.04952936992049217
+ful,v1,test,2391,12.473752170138889,163,79015,,,1.0,0.9938650306748467,0.0,0.0,1200.0,0.045320508764158704,0.30750828119249174,0.0,0.04326097294688225
+ful,v2,train,19293,100.58272569444445,168,634015,,,1.0,,,,,,,,0.05677591264247894
+ful,v2,validation,2303,12.00684353298611,21,70671,,,1.0,0.0,0.0,0.0008333333333333334,1200.0,0.05609090008631546,0.35272192251103485,0.05660377358490566,0.12165185064077377
+ful,v2,test,2285,11.87486111111111,22,78730,,,1.0,0.0,0.0,0.0,1200.0,0.06214911723612346,0.3718905472636816,0.02127659574468085,0.14324843883514404
+kpo,v1,train,14414,82.29270833333334,451,488520,,,1.0,,,,,,,,0.023169519379734993
+kpo,v1,validation,1760,10.050687934027778,341,58985,,,1.0,0.9912023460410557,0.0,0.0016666666666666668,1200.0,0.058709841485123336,0.3721533258173619,0.06306306306306306,0.025582583621144295
+kpo,v1,test,1845,10.538685980902779,354,62285,,,1.0,0.980225988700565,0.0,0.004166666666666667,1200.0,0.05917957774745123,0.38243934283538245,0.039603960396039604,0.027645941823720932
+kpo,v2,train,14416,82.30944444444444,369,498564,,,1.0,,,,,,,,0.029201148077845573
+kpo,v2,validation,1803,10.294301215277779,46,54458,,,1.0,0.0,0.0005546311702717693,0.0008333333333333334,1200.0,0.06030335304271182,0.3908001546192501,0.023809523809523808,0.06377407163381577
+kpo,v2,test,1800,10.27833984375,46,56768,,,1.0,0.0,0.0,0.005,1200.0,0.06521279594137543,0.4113870381586917,0.07339449541284404,0.14331279695034027
+lin,v1,train,14399,71.91595052083333,95,394118,0.5727481075074657,0.547927270512892,0.0,,,,,,,,0.10748428106307983
+lin,v1,validation,1844,9.28003689236111,73,50371,0.5591106290672451,0.5289750054594906,0.0,0.9863013698630136,0.0005422993492407809,0.0008333333333333334,1200.0,0.03460324393003911,0.3304277316034867,0.0,0.09772583097219467
+lin,v1,test,1866,9.25636501736111,75,50716,0.5927116827438371,0.5765044561873964,0.0,0.9866666666666667,0.0,0.0025,1200.0,0.03213975865604543,0.3206555999159487,0.020833333333333332,0.11023253947496414
+lin,v2,train,14523,72.300859375,74,400538,0.5739172347311161,0.5502274440877021,0.0,,,,,,,,0.13356807827949524
+lin,v2,validation,1778,9.056223958333334,13,43996,0.5849268841394826,0.5483907627966179,0.0,0.0,0.0005624296962879641,0.0016666666666666668,1200.0,0.05202745704154923,0.39730290456431533,0.058823529411764705,0.16758273541927338
+lin,v2,test,1808,9.095274522569444,10,50671,0.558075221238938,0.5391052081072013,0.0,0.0,0.0,0.0,1200.0,0.03664818140553768,0.356096461573135,0.0,0.1827891767024994
+lug,v1,train,5455,37.29111979166667,334,157819,0.46815994076268047,0.4730248690512063,0.009715857011915673,,,,,,,,0.011130510829389095
+lug,v1,validation,664,4.578490668402778,252,19557,0.43636363636363634,0.4487443392342528,0.006024096385542169,0.9801587301587301,0.0030120481927710845,0.0030120481927710845,664.0,0.07761926675870533,0.2812623274161736,0.0,0.012599829584360123
+lug,v1,test,638,4.267013888888889,232,18025,0.4675118858954041,0.4672960215778827,0.0109717868338558,0.9956896551724138,0.0,0.001567398119122257,638.0,0.07889042995839113,0.2830070803831737,0.06666666666666667,0.015772808343172073
+lug,v2,train,5366,36.706796875,270,158181,0.4707877874104787,0.4699491358214501,0.011181513231457324,,,,,,,,0.01401989720761776
+lug,v2,validation,692,4.716343315972222,35,18292,0.47819767441860467,0.47097661623108666,0.005780346820809248,0.0,0.002890173410404624,0.002890173410404624,692.0,0.09206210365186968,0.3210633946830266,0.0,0.10228798538446426
+lug,v2,test,699,4.713488498263889,35,18928,0.40772532188841204,0.47004437869822485,0.0,0.0,0.0,0.0,699.0,0.09335376162299239,0.3146826162454177,0.0,0.09920287132263184
+mas,v1,train,6867,39.27149305555555,351,161439,0.33032143386165014,0.3047050084355249,0.01689238386486093,,,,,,,,0.01430696900933981
+mas,v1,validation,849,4.901964518229167,255,20511,0.34725536992840095,0.3144225546832568,0.012956419316843345,0.984313725490196,0.002355712603062426,0.002355712603062426,849.0,0.1249573399639218,0.35071707953063885,0.03333333333333333,0.017459811642766
+mas,v1,test,858,4.8661056857638885,251,20495,0.3392434988179669,0.3071456770627714,0.013986013986013986,0.9721115537848606,0.0011655011655011655,0.0011655011655011655,858.0,0.11959014393754575,0.3418563405261602,0.0,0.023704849183559418
+mas,v2,train,6856,39.233433159722225,290,168449,0.3370953370953371,0.30579934087424704,0.01779463243873979,,,,,,,,0.016872040927410126
+mas,v2,validation,857,4.9021435546875,36,19731,0.28180737217598095,0.30726544034528824,0.01866977829638273,0.0,0.002333722287047841,0.002333722287047841,857.0,0.1188991941614718,0.3511413307430792,0.03333333333333333,0.12520156800746918
+mas,v2,test,861,4.903983289930555,36,14265,0.35,0.305768284971545,0.0011614401858304297,0.0,0.0,0.0,861.0,0.14945671223273746,0.371341695172938,0.0,0.09882112592458725
+mlg,v1,train,18526,94.72736111111111,242,605040,,,1.0,,,,,,,,0.06603185087442398
+mlg,v1,validation,2270,11.653570963541666,169,74099,,,1.0,0.9704142011834319,0.0004405286343612335,0.005,1200.0,0.022132552396118707,0.23057644110275688,0.0,0.053831957280635834
+mlg,v1,test,2290,11.67803927951389,172,74563,,,1.0,0.9767441860465116,0.0004366812227074236,0.005,1200.0,0.02148518702305433,0.22957198443579765,0.02631578947368421,0.061045803129673004
+mlg,v2,train,18467,94.460234375,199,605934,,,1.0,,,,,,,,0.08040688931941986
+mlg,v2,validation,2316,11.793525390625,25,66042,,,1.0,0.0,0.0,0.0008333333333333334,1200.0,0.023273068653281246,0.23034076015727392,0.0,0.16099101305007935
+mlg,v2,test,2303,11.805212673611111,27,81726,,,1.0,0.0,0.0,0.0,1200.0,0.03032082813302988,0.30208475269110235,0.058823529411764705,0.1086939200758934
+nyn,v1,train,6783,40.8146875,360,158028,0.4589414713253723,0.4581529855468651,0.0,,,,,,,,0.012956014834344387
+nyn,v1,validation,831,5.10689453125,256,19815,0.4091456077015644,0.41635124905374715,0.0,0.984375,0.0036101083032490976,0.00842358604091456,831.0,0.0880141307090588,0.3157693044644568,0.0,0.020511141046881676
+nyn,v1,test,858,5.162268337673611,261,20127,0.44405594405594406,0.4407015451880558,0.0,0.9846743295019157,0.006993006993006993,0.009324009324009324,858.0,0.08729567247975356,0.31428035982008995,0.03125,0.013053505681455135
+nyn,v2,train,6752,40.78659722222222,293,161490,0.44920023696682465,0.45195368134249797,0.0,,,,,,,,0.01649364084005356
+nyn,v2,validation,861,5.145512152777778,37,17439,0.48664343786295006,0.4542691668100235,0.0,0.0,0.0011614401858304297,0.003484320557491289,861.0,0.10981134239348586,0.3577055140375682,0.0,0.1174653023481369
+nyn,v2,test,859,5.151739908854167,37,19041,0.4447031431897555,0.45233968804159447,0.0,0.0,0.004656577415599534,0.005820721769499418,859.0,0.08570978414999213,0.31141938078005627,0.0,0.09565091878175735
+orm,v1,train,38185,189.93916666666667,359,1201384,0.5558727248919733,0.5585940881516651,0.0,,,,,,,,0.00844616536051035
+orm,v1,validation,3078,14.45927734375,23,98382,0.6624431448992852,0.6772783639283608,0.0,0.0,0.0003248862897985705,0.004166666666666667,1200.0,0.05152365270069728,0.30663674254568773,0.0,0.10001050680875778
+orm,v1,test,3782,18.25799479166667,24,114528,0.48228450555261765,0.5003056021234982,0.0,0.0,0.00026441036488630354,0.0,1200.0,0.05091331377479743,0.3192214111922141,0.0,0.08316301554441452
+orm,v2,train,40141,198.73305555555555,361,1260546,0.5594529284272938,0.562407877221458,0.0,,,,,,,,0.008072424679994583
+orm,v2,validation,2473,11.936488715277777,23,78189,0.545086938940558,0.5611019452864213,0.0,0.0,0.001213101496158512,0.005833333333333334,1200.0,0.08191689368069677,0.4212016838882511,0.0,0.12159649282693863
+orm,v2,test,2431,11.986879340277778,22,75559,0.5281777046482928,0.5585568893182811,0.0,0.0,0.0016454134101192926,0.0016666666666666668,1200.0,0.05724003758652179,0.3209653587259892,0.0,0.0950242355465889
+sid,v1,train,38752,192.68970486111112,468,851037,0.6530649162704155,0.6554629168594187,0.001445086705202312,,,,,,,,0.010550620965659618
+sid,v1,validation,3091,15.339047309027778,26,72341,0.7007440957618893,0.6678370495293126,0.0,0.0,0.00032351989647363315,0.0008333333333333334,1200.0,0.15139409187044692,0.4771844660194175,0.05714285714285714,0.11289771646261215
+sid,v1,test,3561,17.805811631944444,25,85653,0.6515023869699522,0.6309294478885736,0.0,0.0,0.0,0.0,1200.0,0.10330052654314502,0.4087040226973351,0.0,0.09092046320438385
+sid,v2,train,40547,201.95128472222223,463,899262,0.6554296016398706,0.6538247330584275,0.001381113275951365,,,,,,,,0.010066764429211617
+sid,v2,validation,2444,11.909014756944444,28,56159,0.632569558101473,0.6557274880250716,0.0,0.0,0.0,0.0,1200.0,0.11123773571466729,0.3893306346682088,0.0,0.12276582419872284
+sid,v2,test,2413,11.974242621527777,28,53610,0.6929133858267716,0.6601193807125536,0.0,0.0,0.0024865312888520514,0.005833333333333334,1200.0,0.13633650438351053,0.41805638872225553,0.027777777777777776,0.13035981357097626
+sna,v1,train,14109,79.74210069444445,157,329632,0.5794173931533064,0.6130867148820502,0.0,,,,,,,,0.08992531150579453
+sna,v1,validation,1727,9.712536892361111,100,39788,0.569195136074117,0.6081984517945109,0.0,0.97,0.0005790387955993051,0.0,1200.0,0.07736000804262591,0.3101424979044426,0.0,0.09132098406553268
+sna,v1,test,1749,9.937477213541667,115,40669,0.5740423098913665,0.6099486095060119,0.0,0.9217391304347826,0.0,0.0,1200.0,0.07747916103174408,0.3151138716356108,0.02702702702702703,0.08977515995502472
+sna,v2,train,14079,79.50754340277778,131,330220,0.5837772569074509,0.6139391920537823,0.0,,,,,,,,0.11225946247577667
+sna,v2,validation,1744,9.948648003472222,20,37228,0.5504587155963303,0.6115020951971635,0.0,0.0,0.0,0.0,1200.0,0.08678951326958204,0.34484763898581067,0.02631578947368421,0.19890795648097992
+sna,v2,test,1762,9.935922309027777,17,42641,0.55788876276958,0.6003142515419432,0.0,0.0,0.0005675368898978433,0.0,1200.0,0.08637227081916465,0.3590223608944358,0.0,0.16191913187503815
+sog,v1,train,6219,39.96907118055555,316,163741,0.44694637939413573,0.47834423571336093,0.00739668757034893,,,,,,,,0.017983250319957733
+sog,v1,validation,841,5.337380642361111,230,22386,0.4506024096385542,0.47965816628028546,0.013079667063020214,0.9826086956521739,0.0,0.0047562425683709865,841.0,0.1043062628428482,0.3575897267039842,0.1,0.017150631174445152
+sog,v1,test,817,5.236682942708334,233,21381,0.4591584158415842,0.4848095849379546,0.011015911872705019,0.9785407725321889,0.0,0.0024479804161566705,817.0,0.10055656891632758,0.34196277495769883,0.03571428571428571,0.023051394149661064
+sog,v2,train,6298,40.441796875,260,165514,0.43870346598202825,0.47912518761295064,0.010479517307081613,,,,,,,,0.02159268967807293
+sog,v2,validation,786,5.056221788194445,33,21120,0.5407124681933843,0.47684659090909093,0.0,0.0,0.0,0.0,786.0,0.1324810606060606,0.40284508363295296,0.0,0.11624787002801895
+sog,v2,test,793,5.045118272569445,32,20874,0.43505674653215637,0.4816518156558398,0.0,0.0,0.0,0.0,793.0,0.08187218549391588,0.3295672564485184,0.0,0.12086080759763718
+tir,v1,train,40856,181.92131944444444,384,1043459,0.5645927158801645,0.563363773756324,0.0,,,,,,,,0.00946103036403656
+tir,v1,validation,4425,16.804866536458334,35,111063,0.6061016949152542,0.6228626995489047,0.0,0.0,0.0047457627118644066,0.004166666666666667,1200.0,0.08042282308239468,0.35479637811119374,0.0034129692832764505,0.07899640500545502
+tir,v1,test,5034,20.260447048611113,38,140668,0.4185538339292809,0.41271646714249155,0.0,0.0,0.020460866110448946,0.0175,1200.0,0.10493502431256575,0.44185277088502894,0.013468013468013467,0.07498017698526382
+tir,v2,train,44945,194.84682291666667,408,1158049,0.5516965179664034,0.5518730209170769,0.0,,,,,,,,0.008833416737616062
+tir,v2,validation,2686,12.05203125,25,64675,0.5785554728220402,0.5562118283726324,0.0,0.0,0.0,0.0,1200.0,0.11264012369540008,0.38292737680996747,0.0,0.09265342354774475
+tir,v2,test,2684,12.087771267361111,24,72466,0.5611028315946349,0.5521347942483371,0.0,0.0,0.0007451564828614009,0.0008333333333333334,1200.0,0.14121105069963846,0.45767890390031984,0.0035587188612099642,0.11601842194795609
+wal,v1,train,41833,197.31319444444443,392,936469,0.6358138311859058,0.6180418145181528,0.0,,,,,,,,0.013443926349282265
+wal,v1,validation,2444,9.606139322916667,17,52689,0.9979541734860884,0.9981779878152935,0.0,0.0,0.0,0.0008333333333333334,1200.0,0.06299227542750858,0.2424287743961714,0.0,0.12454313784837723
+wal,v1,test,3014,12.404654947916667,18,64596,0.9996682149966821,0.9996594216360146,0.0,0.0,0.0,0.0008333333333333334,1200.0,0.06961731376555824,0.27046883755681467,0.058823529411764705,0.09907053411006927
+wal,v2,train,42258,195.13776041666668,381,945007,0.6795399687633111,0.6606818785469314,0.0,,,,,,,,0.013593802228569984
+wal,v2,validation,2572,12.164142795138888,23,54404,0.6741835147744946,0.6609440482317477,0.0,0.0,0.0007776049766718507,0.0016666666666666668,1200.0,0.0886515697375193,0.31076342055570544,0.0,0.13607849180698395
+wal,v2,test,2461,12.022080078125,23,54343,0.650142218610321,0.6557790331781462,0.0,0.0,0.0,0.0008333333333333334,1200.0,0.10098816774929614,0.32449754215667975,0.0,0.1252458244562149

data/ASR_v2/metadata/split_headline.csv

@@ -0,0 +1,20 @@
+language,v1_speaker_leak_max,v2_speaker_leak_max,v1_exact_sentence_overlap_max,v2_exact_sentence_overlap_max,v1_near_dup_max,v2_near_dup_max,v1_max_speaker_share,v2_max_speaker_share,v2_unseen_char_max,gender_available
+ach,0.9896907216494846,0.0,0.005681818181818182,0.015384615384615385,0.005681818181818182,0.015384615384615385,0.026069996878504753,0.13066771626472473,0.0,True
+aka,1.0,0.0,0.0,0.0,0.009795191451469279,0.015,0.04804767668247223,0.14713680744171143,0.058823529411764705,False
+amh,0.0,0.0,0.0011695906432748538,0.003709810387469085,0.008333333333333333,0.0033333333333333335,0.1053454726934433,0.12144400179386139,0.007575757575757576,True
+dag,0.9794721407624634,0.0,0.0017142857142857142,0.0005605381165919282,0.005,0.0025,0.015343086794018745,0.12565946578979492,0.05,False
+dga,1.0,0.0,0.19070258848388802,0.002109704641350211,0.19333333333333333,0.0075,0.04143372178077698,0.1290062516927719,0.030303030303030304,False
+ewe,0.9975062344139651,0.0,0.0,0.0,0.0008333333333333334,0.0,0.025948595255613327,0.1145356222987175,0.019230769230769232,False
+ful,0.9938650306748467,0.0,0.0,0.0,0.0,0.0008333333333333334,0.04952936992049217,0.14324843883514404,0.05660377358490566,False
+kpo,0.9912023460410557,0.0,0.0,0.0005546311702717693,0.004166666666666667,0.005,0.027645941823720932,0.14331279695034027,0.07339449541284404,False
+lin,0.9866666666666667,0.0,0.0005422993492407809,0.0005624296962879641,0.0025,0.0016666666666666668,0.11023253947496414,0.1827891767024994,0.058823529411764705,True
+lug,0.9956896551724138,0.0,0.0030120481927710845,0.002890173410404624,0.0030120481927710845,0.002890173410404624,0.015772808343172073,0.10228798538446426,0.0,True
+mas,0.984313725490196,0.0,0.002355712603062426,0.002333722287047841,0.002355712603062426,0.002333722287047841,0.023704849183559418,0.12520156800746918,0.03333333333333333,True
+mlg,0.9767441860465116,0.0,0.0004405286343612335,0.0,0.005,0.0008333333333333334,0.061045803129673004,0.16099101305007935,0.058823529411764705,False
+nyn,0.9846743295019157,0.0,0.006993006993006993,0.004656577415599534,0.009324009324009324,0.005820721769499418,0.020511141046881676,0.1174653023481369,0.0,True
+orm,0.0,0.0,0.0003248862897985705,0.0016454134101192926,0.004166666666666667,0.005833333333333334,0.10001050680875778,0.12159649282693863,0.0,True
+sid,0.0,0.0,0.00032351989647363315,0.0024865312888520514,0.0008333333333333334,0.005833333333333334,0.11289771646261215,0.13035981357097626,0.027777777777777776,True
+sna,0.97,0.0,0.0005790387955993051,0.0005675368898978433,0.0,0.0,0.09132098406553268,0.19890795648097992,0.02631578947368421,True
+sog,0.9826086956521739,0.0,0.0,0.0,0.0047562425683709865,0.0,0.023051394149661064,0.12086080759763718,0.0,True
+tir,0.0,0.0,0.020460866110448946,0.0007451564828614009,0.0175,0.0008333333333333334,0.07899640500545502,0.11601842194795609,0.0035587188612099642,True
+wal,0.0,0.0,0.0,0.0007776049766718507,0.0008333333333333334,0.0016666666666666668,0.12454313784837723,0.13607849180698395,0.0,True