Datasets:
One dataset, two tabs: sentences with take counts, and takes
Browse files- README.md +16 -8
- data/train-20261116T122653Z-r2cg.parquet +2 -2
- meta/schema.json +12 -4
- meta/train.jsonl +3 -3
- sentences/train.parquet +3 -0
README.md
CHANGED
|
@@ -15,12 +15,17 @@ tags:
|
|
| 15 |
- intent
|
| 16 |
- english
|
| 17 |
size_categories:
|
| 18 |
-
- n<
|
| 19 |
configs:
|
| 20 |
-
- config_name:
|
|
|
|
| 21 |
data_files:
|
| 22 |
- split: train
|
| 23 |
path: data/train-*.parquet
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
---
|
| 25 |
# Scripted Voice Notes
|
| 26 |
|
|
@@ -34,11 +39,13 @@ One speaker recording voice notes from scripts: the text is synthetic, the voice
|
|
| 34 |
|
| 35 |
**Website: [hebrish.app](https://hebrish.app)**: get the Hebrish keyboard for Android and the desktop apps, see the results, and find every model, dataset and demo in one place.
|
| 36 |
|
| 37 |
-
**
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
|
|
|
|
|
|
| 42 |
|
| 43 |
**Speaker:** Daniel Rosehill, the dataset's author, who was born in Ireland and speaks English with an **Irish accent**, and who lives in Jerusalem. Every recording is his voice. **Deliberate variation:** he sometimes changes his speaking speed on purpose (fast and rushed, or slow) and sometimes says a word the way people commonly mispronounce it, to make the models robust to real speech. The transcript is always the sentence as written, so a take whose speed or pronunciation sounds off against its card is usually intentional, not a recording error.
|
| 44 |
|
|
@@ -47,6 +54,7 @@ All audio was recorded on a **OnePlus Nord 3 5G** phone using its **internal mic
|
|
| 47 |
## Layout
|
| 48 |
|
| 49 |
```
|
|
|
|
| 50 |
data/train-*.parquet one row per take: `audio` (the FLAC or WAV file, embedded) and every column
|
| 51 |
meta/train.jsonl index: the same rows without audio, plus `shard`, `audio_path` and the audio's byte range
|
| 52 |
meta/schema.json column types
|
|
@@ -60,7 +68,7 @@ The `test` split is **term-disjoint**: a held-out term never appears in `train`.
|
|
| 60 |
|
| 61 |
## Columns
|
| 62 |
|
| 63 |
-
The columns are the same as in the main dataset: `text` (ground truth, punctuated and cased like
|
| 64 |
Parakeet v2 output), `text_he` (the same transcript with its Hebrew terms in Hebrew script, for
|
| 65 |
Hebrew-capable models), `categories`, `target_terms`, `prompt_source`, `prompt_origin` (manual / dataset / generated), `edited`, `duration_s`,
|
| 66 |
`environment`, `conditions`, `device`, `speaker`, `speaker_gender`, `contributor`, `hebrew_pronunciation`, `variant_groups`, `variant_spelling` and `recorded_at`. The main dataset's card describes them.
|
|
|
|
| 15 |
- intent
|
| 16 |
- english
|
| 17 |
size_categories:
|
| 18 |
+
- 1K<n<10K
|
| 19 |
configs:
|
| 20 |
+
- config_name: takes
|
| 21 |
+
default: true
|
| 22 |
data_files:
|
| 23 |
- split: train
|
| 24 |
path: data/train-*.parquet
|
| 25 |
+
- config_name: sentences
|
| 26 |
+
data_files:
|
| 27 |
+
- split: train
|
| 28 |
+
path: sentences/train.parquet
|
| 29 |
---
|
| 30 |
# Scripted Voice Notes
|
| 31 |
|
|
|
|
| 39 |
|
| 40 |
**Website: [hebrish.app](https://hebrish.app)**: get the Hebrish keyboard for Android and the desktop apps, see the results, and find every model, dataset and demo in one place.
|
| 41 |
|
| 42 |
+
**Two tabs.** `sentences` lists every ground-truth sentence with a `takes` count, so
|
| 43 |
+
`takes = 0` shows what has not been recorded. `takes` has one row per recording with its audio, joined on `sentence_id`; a
|
| 44 |
+
sentence can have several takes. `load_dataset("danielrosehill/Scripted-Voice-Notes", "sentences")` reads the text; the default config reads the takes.
|
| 45 |
+
|
| 46 |
+
Takes come from [`danielrosehill/Read-Aloud-Voice-Sentences`](https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences), the full dataset, with the same ids and
|
| 47 |
+
train/test split, by `scripts/sync_subsets.py` in the project repository. Do not edit this dataset directly. New takes append and
|
| 48 |
+
only the sentence counts are rewritten.
|
| 49 |
|
| 50 |
**Speaker:** Daniel Rosehill, the dataset's author, who was born in Ireland and speaks English with an **Irish accent**, and who lives in Jerusalem. Every recording is his voice. **Deliberate variation:** he sometimes changes his speaking speed on purpose (fast and rushed, or slow) and sometimes says a word the way people commonly mispronounce it, to make the models robust to real speech. The transcript is always the sentence as written, so a take whose speed or pronunciation sounds off against its card is usually intentional, not a recording error.
|
| 51 |
|
|
|
|
| 54 |
## Layout
|
| 55 |
|
| 56 |
```
|
| 57 |
+
sentences/train.parquet one row per sentence: `id`, `text`, `takes`, `prompt_ids`, `served` and the sentence columns
|
| 58 |
data/train-*.parquet one row per take: `audio` (the FLAC or WAV file, embedded) and every column
|
| 59 |
meta/train.jsonl index: the same rows without audio, plus `shard`, `audio_path` and the audio's byte range
|
| 60 |
meta/schema.json column types
|
|
|
|
| 68 |
|
| 69 |
## Columns
|
| 70 |
|
| 71 |
+
A take's `sentence_id` is its row's `id` in `sentences`. A sentence with `served: false` is no longer offered by the app; it is kept because a take used it. The columns are the same as in the main dataset: `text` (ground truth, punctuated and cased like
|
| 72 |
Parakeet v2 output), `text_he` (the same transcript with its Hebrew terms in Hebrew script, for
|
| 73 |
Hebrew-capable models), `categories`, `target_terms`, `prompt_source`, `prompt_origin` (manual / dataset / generated), `edited`, `duration_s`,
|
| 74 |
`environment`, `conditions`, `device`, `speaker`, `speaker_gender`, `contributor`, `hebrew_pronunciation`, `variant_groups`, `variant_spelling` and `recorded_at`. The main dataset's card describes them.
|
data/train-20261116T122653Z-r2cg.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee1808055b46752fff897a9b550b6983d86ae9f9a9338eeeaf68a50aab3e1afe
|
| 3 |
+
size 1839111
|
meta/schema.json
CHANGED
|
@@ -227,7 +227,7 @@
|
|
| 227 |
],
|
| 228 |
[
|
| 229 |
"written_form_uncertain",
|
| 230 |
-
"
|
| 231 |
],
|
| 232 |
[
|
| 233 |
"session_id",
|
|
@@ -295,15 +295,15 @@
|
|
| 295 |
],
|
| 296 |
[
|
| 297 |
"latitude",
|
| 298 |
-
"
|
| 299 |
],
|
| 300 |
[
|
| 301 |
"longitude",
|
| 302 |
-
"
|
| 303 |
],
|
| 304 |
[
|
| 305 |
"location_accuracy_m",
|
| 306 |
-
"
|
| 307 |
],
|
| 308 |
[
|
| 309 |
"text_full",
|
|
@@ -312,6 +312,14 @@
|
|
| 312 |
[
|
| 313 |
"fallback_ok",
|
| 314 |
"bool"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 315 |
]
|
| 316 |
]
|
| 317 |
}
|
|
|
|
| 227 |
],
|
| 228 |
[
|
| 229 |
"written_form_uncertain",
|
| 230 |
+
"bool"
|
| 231 |
],
|
| 232 |
[
|
| 233 |
"session_id",
|
|
|
|
| 295 |
],
|
| 296 |
[
|
| 297 |
"latitude",
|
| 298 |
+
"float64"
|
| 299 |
],
|
| 300 |
[
|
| 301 |
"longitude",
|
| 302 |
+
"float64"
|
| 303 |
],
|
| 304 |
[
|
| 305 |
"location_accuracy_m",
|
| 306 |
+
"float64"
|
| 307 |
],
|
| 308 |
[
|
| 309 |
"text_full",
|
|
|
|
| 312 |
[
|
| 313 |
"fallback_ok",
|
| 314 |
"bool"
|
| 315 |
+
],
|
| 316 |
+
[
|
| 317 |
+
"sentence_id",
|
| 318 |
+
"string"
|
| 319 |
+
],
|
| 320 |
+
[
|
| 321 |
+
"pronunciation",
|
| 322 |
+
"json"
|
| 323 |
]
|
| 324 |
]
|
| 325 |
}
|
meta/train.jsonl
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
-
{"id": "20261116T075406Z_c9u2va", "text": "Dear Mr. Ryan, I am writing to share news of my move to Copenhagen. My new address will be 12 Strand Street. Please update your records accordingly. I would be grateful if you could confirm receipt of this change. Yours faithfully, Sarah", "text_full": null, "fallback_ok": null, "text_he": "Dear Mr. Ryan, I am writing to share news of my move to Copenhagen. My new address will be 12 Strand Street. Please update your records accordingly. I would be grateful if you could confirm receipt of this change. Yours faithfully, Sarah", "duration_s": 15.794, "categories": ["voice-notes"], "target_terms": [], "prompt_id": "voice-notes_166a6c2339", "prompt_source": "deepseek-v4.1-flash", "prompt_origin": "dataset", "mode": "bank", "edited": false, "device": "OnePlus Nord 3 5G", "microphone": "internal", "mic_link": null, "bt_codec": null, "mic_name": "CPH2493", "mic_usb_vendor_id": null, "mic_usb_product_id": null, "mic_type": "builtin_mic", "mic_chain": null, "activity": "seated", "environment": "quiet-indoor", "conditions": [], "noise_level": null, "background_speech_languages": [], "background_speech_kinds": [], "noise_flags": [], "clipped": false, "notes":
|
| 2 |
-
{"id": "20261116T080030Z_y77f8e", "text": "Hi all, Sorry for the delay on the report. I had to wait for Daniel to send the final numbers. It is ready now and I will email it today. Thanks for your patience. Thanks again, Ruth", "text_full": null, "fallback_ok": null, "text_he": "Hi all, Sorry for the delay on the report. I had to wait for Daniel to send the final numbers. It is ready now and I will email it today. Thanks for your patience. Thanks again, Ruth", "duration_s": 10.357, "categories": ["voice-notes"], "target_terms": [], "prompt_id": "voice-notes_2a68807bb8", "prompt_source": "deepseek-v4.1-flash", "prompt_origin": "dataset", "mode": "bank", "edited": false, "device": "OnePlus Nord 3 5G", "microphone": "internal", "mic_link": null, "bt_codec": null, "mic_name": "CPH2493", "mic_usb_vendor_id": null, "mic_usb_product_id": null, "mic_type": "builtin_mic", "mic_chain": null, "activity": "seated", "environment": "quiet-indoor", "conditions": [], "noise_level": null, "background_speech_languages": [], "background_speech_kinds": [], "noise_flags": [], "clipped": false, "notes":
|
| 3 |
-
{"id": "20261116T080345Z_nuxc2b", "text": "Hello Avi, I would like to invite you to dinner at my home. We are planning a small gathering on Saturday at 7 PM. Please let me know if you can make it. It would be lovely to see you. I hope the date works for you. Best regards, Owen", "text_full": null, "fallback_ok": null, "text_he": "Hello Avi, I would like to invite you to dinner at my home. We are planning a small gathering on Saturday at 7 PM. Please let me know if you can make it. It would be lovely to see you. I hope the date works for you. Best regards, Owen", "duration_s": 13.805, "categories": ["voice-notes"], "target_terms": [], "prompt_id": "voice-notes_e3a538ce05", "prompt_source": "deepseek-v4.1-flash", "prompt_origin": "dataset", "mode": "bank", "edited": false, "device": "OnePlus Nord 3 5G", "microphone": "internal", "mic_link": null, "bt_codec": null, "mic_name": "CPH2493", "mic_usb_vendor_id": null, "mic_usb_product_id": null, "mic_type": "builtin_mic", "mic_chain": null, "activity": "seated", "environment": "quiet-indoor", "conditions": [], "noise_level": null, "background_speech_languages": [], "background_speech_kinds": [], "noise_flags": [], "clipped": false, "notes":
|
|
|
|
| 1 |
+
{"id": "20261116T075406Z_c9u2va", "text": "Dear Mr. Ryan, I am writing to share news of my move to Copenhagen. My new address will be 12 Strand Street. Please update your records accordingly. I would be grateful if you could confirm receipt of this change. Yours faithfully, Sarah", "text_full": null, "fallback_ok": null, "text_he": "Dear Mr. Ryan, I am writing to share news of my move to Copenhagen. My new address will be 12 Strand Street. Please update your records accordingly. I would be grateful if you could confirm receipt of this change. Yours faithfully, Sarah", "duration_s": 15.794, "categories": ["voice-notes"], "target_terms": [], "prompt_id": "voice-notes_166a6c2339", "prompt_source": "deepseek-v4.1-flash", "prompt_origin": "dataset", "mode": "bank", "edited": false, "device": "OnePlus Nord 3 5G", "microphone": "internal", "mic_link": null, "bt_codec": null, "mic_name": "CPH2493", "mic_usb_vendor_id": null, "mic_usb_product_id": null, "mic_type": "builtin_mic", "mic_chain": null, "activity": "seated", "environment": "quiet-indoor", "conditions": [], "noise_level": null, "background_speech_languages": [], "background_speech_kinds": [], "noise_flags": [], "clipped": false, "notes": null, "hebrew_pronunciation": null, "pronunciation_variants": [], "variant_groups": [], "variant_spelling": null, "task": "vn-emails", "priority_word": null, "spoken_text": "Dear Mr. Ryan,\n\nI am writing to share news of my move to Copenhagen. My new address will be 12 Strand Street. Please update your records accordingly. I would be grateful if you could confirm receipt of this change.\n\nYours faithfully, Sarah", "aside_text": null, "aside_position": null, "verbatim_text": null, "prompt_categories": null, "source_prompt_id": null, "prompt_kind": "email", "prompt_note": "formal, sharing news about a move", "ui_note": null, "term_language": null, "paragraph_text": "Dear Mr. Ryan,\n\nI am writing to share news of my move to Copenhagen. My new address will be 12 Strand Street. Please update your records accordingly. I would be grateful if you could confirm receipt of this change.\n\nYours faithfully, Sarah", "source_url": null, "model_hypothesis": null, "asr_model": null, "check_verdict": null, "check_errors": null, "check_error_types": null, "speaker": "Daniel Rosehill", "speaker_id": "8992", "contributor": "Daniel Rosehill", "speaker_gender": "male", "sample_rate": 48000, "bit_depth": 16, "audio_format": "flac", "native_rate": 48000, "capture_mode": "stt", "tts_eligible": true, "tts_split": null, "transcript_status": "script", "prompt_set": null, "topic": null, "style": null, "style_instruction": null, "written_form_uncertain": null, "session_id": "20261116T075312Z_m1moqg", "room_tone_id": null, "room": null, "location": null, "location_name": null, "location_note": null, "latitude": null, "longitude": null, "location_accuracy_m": null, "lead_trim_s": 0, "peak_dbfs": -4.3, "floor_dbfs": -49.7, "speech_dbfs": -21.9, "snr_db": 27.8, "clipped_samples": 0, "edge_head_dbfs": -39.9, "edge_tail_dbfs": -34.6, "browser_processing": false, "recorded_at": "2026-11-16T07:54:06.569Z", "sentence_id": "p_abcd131352", "pronunciation": [], "shard": "data/train-20261116T122653Z-r2cg.parquet", "audio_path": "audio/2026-11-16/20261116T075406Z_c9u2va.flac", "audio_offset": 39, "audio_length": 659129}
|
| 2 |
+
{"id": "20261116T080030Z_y77f8e", "text": "Hi all, Sorry for the delay on the report. I had to wait for Daniel to send the final numbers. It is ready now and I will email it today. Thanks for your patience. Thanks again, Ruth", "text_full": null, "fallback_ok": null, "text_he": "Hi all, Sorry for the delay on the report. I had to wait for Daniel to send the final numbers. It is ready now and I will email it today. Thanks for your patience. Thanks again, Ruth", "duration_s": 10.357, "categories": ["voice-notes"], "target_terms": [], "prompt_id": "voice-notes_2a68807bb8", "prompt_source": "deepseek-v4.1-flash", "prompt_origin": "dataset", "mode": "bank", "edited": false, "device": "OnePlus Nord 3 5G", "microphone": "internal", "mic_link": null, "bt_codec": null, "mic_name": "CPH2493", "mic_usb_vendor_id": null, "mic_usb_product_id": null, "mic_type": "builtin_mic", "mic_chain": null, "activity": "seated", "environment": "quiet-indoor", "conditions": [], "noise_level": null, "background_speech_languages": [], "background_speech_kinds": [], "noise_flags": [], "clipped": false, "notes": null, "hebrew_pronunciation": null, "pronunciation_variants": [], "variant_groups": [], "variant_spelling": null, "task": "voice-notes", "priority_word": null, "spoken_text": "Hi all,\n\nSorry for the delay on the report. I had to wait for Daniel to send the final numbers. It is ready now and I will email it today. Thanks for your patience.\n\nThanks again, Ruth", "aside_text": null, "aside_position": null, "verbatim_text": null, "prompt_categories": null, "source_prompt_id": null, "prompt_kind": "email", "prompt_note": "brisk, a short apology for a delay", "ui_note": null, "term_language": null, "paragraph_text": "Hi all,\n\nSorry for the delay on the report. I had to wait for Daniel to send the final numbers. It is ready now and I will email it today. Thanks for your patience.\n\nThanks again, Ruth", "source_url": null, "model_hypothesis": null, "asr_model": null, "check_verdict": null, "check_errors": null, "check_error_types": null, "speaker": "Daniel Rosehill", "speaker_id": "8992", "contributor": "Daniel Rosehill", "speaker_gender": "male", "sample_rate": 48000, "bit_depth": 16, "audio_format": "flac", "native_rate": 48000, "capture_mode": "stt", "tts_eligible": true, "tts_split": null, "transcript_status": "script", "prompt_set": null, "topic": null, "style": null, "style_instruction": null, "written_form_uncertain": null, "session_id": "20261116T075804Z_hak4cj", "room_tone_id": null, "room": null, "location": null, "location_name": null, "location_note": null, "latitude": null, "longitude": null, "location_accuracy_m": null, "lead_trim_s": 0, "peak_dbfs": -1.5, "floor_dbfs": -46.7, "speech_dbfs": -16.5, "snr_db": 30.2, "clipped_samples": 0, "edge_head_dbfs": -34.6, "edge_tail_dbfs": -32.4, "browser_processing": false, "recorded_at": "2026-11-16T08:00:30.396Z", "sentence_id": "p_dcae329983", "pronunciation": [], "shard": "data/train-20261116T122653Z-r2cg.parquet", "audio_path": "audio/2026-11-16/20261116T080030Z_y77f8e.flac", "audio_offset": 659172, "audio_length": 500921}
|
| 3 |
+
{"id": "20261116T080345Z_nuxc2b", "text": "Hello Avi, I would like to invite you to dinner at my home. We are planning a small gathering on Saturday at 7 PM. Please let me know if you can make it. It would be lovely to see you. I hope the date works for you. Best regards, Owen", "text_full": null, "fallback_ok": null, "text_he": "Hello Avi, I would like to invite you to dinner at my home. We are planning a small gathering on Saturday at 7 PM. Please let me know if you can make it. It would be lovely to see you. I hope the date works for you. Best regards, Owen", "duration_s": 13.805, "categories": ["voice-notes"], "target_terms": [], "prompt_id": "voice-notes_e3a538ce05", "prompt_source": "deepseek-v4.1-flash", "prompt_origin": "dataset", "mode": "bank", "edited": false, "device": "OnePlus Nord 3 5G", "microphone": "internal", "mic_link": null, "bt_codec": null, "mic_name": "CPH2493", "mic_usb_vendor_id": null, "mic_usb_product_id": null, "mic_type": "builtin_mic", "mic_chain": null, "activity": "seated", "environment": "quiet-indoor", "conditions": [], "noise_level": null, "background_speech_languages": [], "background_speech_kinds": [], "noise_flags": [], "clipped": false, "notes": null, "hebrew_pronunciation": null, "pronunciation_variants": [], "variant_groups": [], "variant_spelling": null, "task": "voice-notes", "priority_word": null, "spoken_text": "Hello Avi,\n\nI would like to invite you to dinner at my home. We are planning a small gathering on Saturday at 7 PM. Please let me know if you can make it.\n\nIt would be lovely to see you. I hope the date works for you.\n\nBest regards, Owen", "aside_text": null, "aside_position": null, "verbatim_text": null, "prompt_categories": null, "source_prompt_id": null, "prompt_kind": "email", "prompt_note": "professional, inviting someone to dinner", "ui_note": null, "term_language": null, "paragraph_text": "Hello Avi,\n\nI would like to invite you to dinner at my home. We are planning a small gathering on Saturday at 7 PM. Please let me know if you can make it.\n\nIt would be lovely to see you. I hope the date works for you.\n\nBest regards, Owen", "source_url": null, "model_hypothesis": null, "asr_model": null, "check_verdict": null, "check_errors": null, "check_error_types": null, "speaker": "Daniel Rosehill", "speaker_id": "8992", "contributor": "Daniel Rosehill", "speaker_gender": "male", "sample_rate": 48000, "bit_depth": 16, "audio_format": "flac", "native_rate": 48000, "capture_mode": "stt", "tts_eligible": true, "tts_split": null, "transcript_status": "script", "prompt_set": null, "topic": null, "style": null, "style_instruction": null, "written_form_uncertain": null, "session_id": "20261116T075804Z_hak4cj", "room_tone_id": null, "room": null, "location": null, "location_name": null, "location_note": null, "latitude": null, "longitude": null, "location_accuracy_m": null, "lead_trim_s": 0, "peak_dbfs": -2.7, "floor_dbfs": -45.1, "speech_dbfs": -15.5, "snr_db": 29.6, "clipped_samples": 0, "edge_head_dbfs": -30.3, "edge_tail_dbfs": -37.7, "browser_processing": false, "recorded_at": "2026-11-16T08:03:45.833Z", "sentence_id": "p_805fa8ee46", "pronunciation": [], "shard": "data/train-20261116T122653Z-r2cg.parquet", "audio_path": "audio/2026-11-16/20261116T080345Z_nuxc2b.flac", "audio_offset": 1160097, "audio_length": 638406}
|
sentences/train.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8538c614b90012e7621b406b8e6a0aa1e6e0378b120ab8d906789a302eafe23
|
| 3 |
+
size 960974
|