Datasets:

Hebrish
/

Dataset-Index

built timestamp[s]	master string	blends dict	models list	datasets list
2026-11-09T00:00:00	danielrosehill/Read-Aloud-Voice-Sentences	{ "cs": { "line": "Jachnun (and the first Shakshuka runs)", "rule": "code-switching rows, Daniel's own English (about 2/3 of the English share), LibriSpeech-PC replay, plus the task tags ['currency', 'long-code-switch']" }, "all": { "line": "Sabich, Shakshuka v1.0+ and the Sabich variants", "rule"...	[ { "id": "danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching", "line": "Jachnun", "version": "PoC-1", "run": "pilot-en30", "released": "2026-10-15T00:00:00", "task": "speech-recognition", "base": null, "trained_on": [ "danielrosehill/Read-Aloud-Voice-Sentences" ...	[ { "id": "danielrosehill/Read-Aloud-Voice-Sentences", "title": "Read-Aloud Voice Sentences (master)", "kind": "audio-takes", "tag": null, "description": "Every take from the capture app; all speech models train on a pinned revision of it, selected by blend. Gated.", "in_blends": [ "cs",...

Hebrish Dataset Index

The map of the Hebrish datasets, across the Hebrish organisation and the danielrosehill account, sorted by kind, and which models were trained on what. Machine-readable: index.json. Generated by scripts/build_dataset_index.py in the training repo; last built 2026-11-09.

114 datasets (110 published; the rest are private or registered tasks whose subset appears once takes are recorded).
28 models.

Sentences and audio

A sentence dataset is text to read. An audio dataset holds the takes. kind in index.json says which a dataset is, account where it lives, and pairs_with links a sentence bank to its audio subset. Master-Sentence-Bank holds every sentence in one file and Read-Aloud-Voice-Sentences every take; the others are topic slices of those two.

How training data works here

Every take is recorded once, into the master dataset danielrosehill/Read-Aloud-Voice-Sentences. The per-topic datasets (code-switching, tech, acronyms, whispered, ...) are views of it, published for browsing and reuse. A speech model trains on the master at a pinned revision, through a blend: a rule for which rows it takes. in_blends on each dataset says which blends include its rows.

Blend	Used by	Rule
`cs`	Jachnun (and the first Shakshuka runs)	code-switching rows, Daniel's own English (about 2/3 of the English share), LibriSpeech-PC replay, plus the task tags ['currency', 'long-code-switch']
`all`	Sabich, Shakshuka v1.0+ and the Sabich variants	every row of the master dataset at the run's revision, including every task tag present then, except ['dropped-articles', 'name-cipher', 'paragraphs', 'weird-voices']; replay about 10%. From Sabich v2.1, priority-word takes (Hebrish/ASR-Error-List) appear 3 times.

Map: supersets and subsets

Each split-out is a subset of a master, and some sit inside a larger split-out. A dataset marked superset has subsets of its own. parent and children in index.json carry the same map, and every card shows it as Subset and Superset badges. Hours per audio dataset and the growth over time: danielrosehill/Audio-Dataset-Index.

Audio

Read-Aloud-Voice-Sentences · superset

Sentences (the text that is read aloud)

Master-Sentence-Bank (private) · superset

Datasets

Sentence datasets (text to read) (11)

Dataset	Account	Access	Pairs with	What it is
AI-Interactions-Sentence-Bank	Hebrish	open	AI-Interactions-Sentences	Sentences dictated to AI agents (text only).
AI-Tech-Vocabulary-Sentence-Bank	Hebrish	gated	Tech-Sentences	AI and tech vocabulary sentence bank (text only).
hebrew-latin-code-switching	Hebrish	gated		Sentences with romanized Hebrew words labelled, for the token classifier.
Homophone-Context-Sentence-Bank	Hebrish	gated	Homophone-Context-Sentences	Homophone-in-context sentence bank (text only).
Linux-Admin-Sentence-Bank	Hebrish	gated	Linux-Admin-Sentences	Linux administration sentence bank (text only).
Master-Sentence-Bank	Hebrish	private		Every sentence written or read for the speech models, one row per unique sentence, categorised, with take counts. Text only.
October-7-Memorial-Sentence-Bank	Hebrish	gated	October-7-Memorial-Sentences	In memory of those who were murdered and who fell on October 7, 2023: three sentences about each person's life, and places and terms connected with that day.
Personal-Vocabulary-Sentence-Bank	Hebrish	gated	Personal-Vocabulary-Sentences	Personal vocabulary sentence bank (text only).
Smart-Home-Sentence-Bank	Hebrish	open	Smart-Home-Sentences	Smart-home sentence bank (text only).
Third-Party-Sentences	Hebrish	private		Sentences from outside text datasets that passed the content screen, each with its source dataset. Text only.
term-context-sentences	danielrosehill	open		Training data for the term-context sentence generator (Granite 4.2 3B).

Audio datasets (read takes) (77)

Dataset	Account	Access	Pairs with	What it is
Acronym-Sentences	Hebrish	open		Natural sentences with acronyms read by one speaker, both spelled out (API, SSH, HTTPS) and said as words (NASA, JSON, YAML), each in its standard written form
AI-Integration-Sentences	Hebrish	open		English sentences naming one item each: apps and services used with an AI assistant: ChatGPT apps, Claude connectors, integration partners and top Zapier apps, read by one speaker, spelled the vendor's way
AI-Model-Sentences	Hebrish	open		English sentences naming one current AI model each (GPT-6.1 Terra, Claude Opus 5.5, Qwen3.8 Max, Kimi K3.1 Pro), read by one speaker, spelled the vendor's way
API-Ops-Sentences	Hebrish	open		English sentences about API work, read by one speaker: HTTP methods in capitals (GET, POST, PUT, PATCH, DELETE, HEAD, OPTIONS), REST API, status codes in digits, auth, webhooks and pull requests, with contrast rows that keep ordinary get, p
Architecture-Practice-Sentences	Hebrish	open		English sentences of professional architecture vocabulary, read by one speaker: reality capture (LiDAR, terrestrial and mobile laser scanning, point clouds, photogrammetry, scan-to-BIM), BIM and design software, urban planning discourse, an
Background-Speech	Hebrish	open		Every take from the main dataset with another voice audible behind the speaker: someone talking in the room, a market crowd, a recorded PA message, TV or radio
Brand-Name-Sentences	Hebrish	gated		Natural sentences with company, product, AI model and open-source project names (Anthropic, OpenRouter, Amazon Bedrock, Qwen, Kubernetes, Nginx) read by one speaker, written in each name's official form
Claude-Dictations	Hebrish	open		One speaker reading his own prompts to an AI coding agent (Claude Code): requests, project directions, feedback and replies, mostly 2 to 9 sentences, for sentence boundaries and punctuation in real dictation
Colloquial-Reduction-Sentences	Hebrish	open		One speaker reading everyday messages with reduced spoken forms (gonna, wanna, gotta, kinda, sorta, 'cause / cuz, lemme, gimme, dunno, outta, lotta, c'mon, ya, tryna, hafta, oughta, shoulda / coulda / woulda)
Community-Voice-Contributions	Hebrish	open		Recordings contributed by anyone, for the speech models.
Currency-Sentences	Hebrish	gated		One speaker saying amounts of money the natural way ("50 shekels", "a dollar fifty", "6 pounds 50")
Data-Tool-Sentences	Hebrish	open		English sentences naming one item each: news data sources and APIs (GDELT and others), scraping frameworks, and vector databases (with a Qdrant top-up), read by one speaker, spelled the vendor's way
Date-Time-Sentences	Hebrish	open		Dates and times said the way people say them and written the way they are typed in an email or a document, read by one speaker: "the second of March at seven thirty p.m." → "March 2 at 7:30PM", Irish "half seven" → "7:30" (never turned into
Dev-Stack-Sentences	Hebrish	open		English sentences naming one item each: libraries, frameworks, CLIs and services from Daniel's own repositories that the bank lacked, read by one speaker, spelled the vendor's way
Disfluent-Read-Sentences	Hebrish	gated		Experimental
Dropped-Article-Sentences	Hebrish	open		One speaker reading terse self-dictated notes (2 or 3 short sentences) in which a later sentence drops "the" before a noun the earlier sentence set up ("I prefer that one
Email-Address-Sentences	Hebrish	open		One speaker saying email addresses aloud inside everyday sentences: role addresses (support, billing, no-reply), first.last, first_last, first-last, initials plus surname, names with digits, handles, plus tags, subdomains, hyphenated and co
English-Hebrew-Mixed-Sentences	Hebrish	gated		The first code-switching set: sentences with recordings (637 WAV files, checked 2026-11-05). Its sentences are also a source of the prompt bank.
Entity-Dense-Sentences	Hebrish	open		One speaker reading sentences that pack three or more entities into one line: company and product names, people, places, amounts, percentages, dates, times and quantities, in the style of results calls, meeting recaps, itineraries, invoices
Expanded-Acronym-Sentences	Hebrish	gated		Chat and business shorthand said as letters by one speaker ("by E O D", "F Y I", "I M O") with the transcript writing it out in full ("by end of day", "For your information", "In my opinion")
False-Start-Sentences	Hebrish	open		Experimental
Fast-Speech-Sentences	Hebrish	open		Robustness data
Fragment-Sentences	Hebrish	open		One speaker reading short and incomplete utterances, one to six words: message replies ("Not yet."), single words ("Apparently."), notes to self ("Dentist, Thursday morning."), list items, headings, trailing clauses ("Unless it rains.") and
Functional-GI-Sentences	Hebrish	open		English sentences of functional gastrointestinal vocabulary, read by one speaker describing his own symptoms, tests and treatment after gallbladder removal: conditions and symptoms (functional dyspepsia, gastroparesis, bile acid diarrhea, a
GitHub-Trending-Sentences	Hebrish	open		English sentences naming one item each: open-source projects trending on GitHub, probed periodically, read by one speaker, spelled the vendor's way
Hard-ASR-Sentences	Hebrish	gated		English sentences speech recognition finds hard, read by one speaker: Harvard sentences (IEEE 1969, public domain), tongue twisters, homographs, garden-path sentences, rare words, dense numbers and units, unusual proper nouns
Hardware-Sentences	Hebrish	open		Short workshop and DIY sentences with hardware vocabulary read by one speaker: electronic components (capacitors, resistors, transistors, MOSFETs, regulators), fasteners and fixings (M3 and M4 screws, standoffs, wall anchors, Torx), connect
Hebrish-TTS-Voice-Recordings	Hebrish	gated
Idiom-Sentences	Hebrish	open		Natural sentences with English idioms that speech recognition tends to spell as a sound-alike ("free rein", not "free reign"; "toe the line", not "tow"; "piqued my interest"), read by one speaker, written in the standard form, plus everyday
Implied-Quote-Sentences	Hebrish	open		Experimental
Irish-English-Sentences	Hebrish	gated		Everyday sentences in Irish English read by one speaker born in Ireland: vocabulary (craic, grand, giving out, press, messages), grammar (I'm after doing it, so it is, ye), Irish-language words used in English (Taoiseach, Dáil, Gardaí, slái
Israeli-Business-Sentences	Hebrish	open		Everyday English sentences naming Israeli businesses read by one speaker: banks and insurers, supermarket and electronics chains, telecoms, food brands, cafes and public services, in the Latin spelling each uses itself (Hazi Hinam, Bank Yah
Israeli-Name-Sentences	Hebrish	gated		English messages and emails that mention an Israeli or Jewish person's name, a first name (Naftali, Yotam, Tzlil, Bat-El) or a surname (Buskila, Ifergan, Weinberg), written in its common Israeli-English spelling, read aloud by one speaker
Israeli-Place-Sentences	Hebrish	open		Everyday English sentences naming one Israeli place each, read by one speaker: second-tier towns, kibbutzim, moshavim and community settlements (chosen from the CBS locality list by population and how often English speakers mention them) an
Israeli-Street-Sentences	Hebrish	open		Everyday English sentences naming one Israeli street each (addresses, directions, errands), read by one speaker: Jerusalem and Tel Aviv first, then Haifa, Be'er Sheva and Ramat Gan, from the data.gov.il street list
Israeli-TV-Sentences	Hebrish	open		Everyday English sentences naming one Israeli TV show each, read by one speaker: long-running satire, sitcoms, dramas, current affairs, reality, talk and kids' shows across Kan, Keshet 12, Reshet 13 and others, by their Hebrew titles romani
Jewish-Religious-Terminology-Sentences	Hebrish	gated		English sentences with Jewish religious vocabulary written in Latin letters: Shabbat and the festivals, prayer and the synagogue, Torah portions, Talmud and halacha, kashrut, and life-cycle events, read aloud by one speaker
Judaica-Sentences	Hebrish	open		Real sentences from traditional Jewish works in English, read by one speaker, Hebrew terms in romanized form: Tanach (Metsudah), Mishnah (Mishnah Yomit, Joshua Kulp; Shraga Silverstein; Sefaria Community Translation; Eighteen Treatises from
Latin-Phrase-Sentences	Hebrish	open		One speaker reading English sentences that use a Latin phrase (de facto, ad hominem, de minimis, bona fide, ad nauseam, per se, quid pro quo), with the phrase in its standard English spelling: lowercase, no italics, no hyphen (`ad hoc`, `po
Long-Code-Switching-Sentences	Hebrish	open		One speaker reading dictation-length English passages with romanized Hebrew, about 25 to 55 seconds in one take
Long-Tech-Sentences	Hebrish	open		One speaker reading dictation-length tech passages, about 25 to 55 seconds in one take
Model-Checks	Hebrish	gated		Takes checked by a speech recognition model while they were recorded, with the model's transcript, the speaker's verdict on it and, once judged, a second, machine verdict
NATO-Alphabet-Sentences	Hebrish	open		Experimental
Noisy-Condition-Sentences	Hebrish	gated		Every take from the main dataset recorded in noise, whatever its category: outdoors, in a noisy room, with traffic, other people talking, background noise or a crying baby
Numerics-Sentences	Hebrish	open		Every kind of number that is not money, said by one speaker in a fixed way and written the way it is typed: the same number said several ways ("thirteen hundred", "one thousand three hundred" → 1,300), decimals, temperatures (written "27 de
Open-Source-Sentences	Hebrish	open		English sentences naming one item each: the biggest open-source and self-hosted projects (awesome-selfhosted, most-starred GitHub repos), read by one speaker, spelled the vendor's way
Paragraph-Dictation-Passages	Hebrish	open		Experimental
Phone-Number-Sentences	Hebrish	open		One speaker dictating phone numbers inside everyday sentences, read digit by digit (oh or zero), with double and triple, or in pairs
Procurement-Sentences	Hebrish	open		Business sentences with purchasing and trade vocabulary read by one speaker: Incoterms 2020 (EXW, FCA, FOB, CIF, DDP, …) with their named place, purchase orders, RFQs, RFPs, invoices, bills of lading, MOQs, lead times, payment terms, freigh
Pronunciation-Variants	Hebrish	gated		Takes in which a word was deliberately said one particular way, from the main dataset, stratified by `variant_category`
Repeated-Word-Sentences	Hebrish	gated		Experimental
Run-On-Sentences	Hebrish	open		One speaker reading run-on dictation in one breath: complete sentences joined by commas, "so" or "and"
Self-Correction-Sentences	Hebrish	open		One speaker correcting himself mid-sentence ("send it Tuesday, no, Wednesday")
Shingle-Sentences	Hebrish	open		English sentences on independent professional practice, read by one speaker: roles in AI and technical work, engagement and pricing terms, outreach and email deliverability, the software around them, and a small group of Israeli business te
Singing-Sentences	Hebrish	open		Experimental
Slow-Speech-Sentences	Hebrish	open		Robustness data
Spelled-Out-Sentences	Hebrish	open		One speaker spelling names, words and codes letter by letter in dictation ("my surname is B R E N N A N"), with the transcript writing the word the letters spell (`My surname is Brennan.`); codes as typed, uppercase letters and digits (`KQ7
Spoken-Aside-Sentences	Hebrish	open		One speaker dictating a sentence and, partway through, saying something to someone in the room ("come in, Hannah")
Spoken-Punctuation-Sentences	Hebrish	open		Experimental
Synthetic-Codeswitching-Voices-Pilot	Hebrish	open		100 synthetic clips (2026-10-28): English sentences with one Hebrew term.
Tanach-Verse-Sentences	Hebrish	open		Real Tanakh verses in English, Metsudah translation via Sefaria (sefaria.org), licence CC-BY, read by one speaker
Thinking-Pause-Sentences	Hebrish	open		Everyday English sentences read by one speaker with a deliberate thinking pause (a second or two of silence) mid-sentence or between two sentences
Web-Address-Sentences	Hebrish	open		One speaker saying web addresses aloud ("hebrish dot app slash benchmarks", "h t t p s colon slash slash", "localhost colon three thousand"), with the transcript writing them as typed: `hebrish.app/benchmarks`, `https://`, `localhost:3000`
Weird-Voice-Sentences	Hebrish	open		Experimental
Whispered-Sentences	Hebrish	open		Experimental robustness data
Yiddish-Sentences	Hebrish	gated		English sentences with Yiddish words written in Latin letters (shul, daven, cholent, mensch, freier, tachles), the way English speakers in Israel use them, read aloud by one speaker
AI-Interactions-Sentences	danielrosehill	open	AI-Interactions-Sentence-Bank	Instructions dictated to AI coding agents (Claude Code, Codex and others) and the git and GitHub workflow around them, read aloud by one speaker: subagents, worktrees, compaction, plan mode, hooks, MCP servers, pull requests, rebases
English-Hebrew-Codeswitching-Sentences	danielrosehill	gated		English sentences with everyday Hebrew words and Israeli names written in Latin letters, read aloud by one English speaker living in Israel
General-English-Sentences	danielrosehill	open		Plain, everyday English sentences (emails, messages, general prose and homophone pairs) read aloud by one speaker, with no Hebrew or Yiddish words, Israeli names or tech vocabulary
Homophone-Context-Sentences	danielrosehill	open	Homophone-Context-Sentence-Bank	English sentences whose context settles which spelling of a homophone is meant (their/there, principal/principle, and 58 more sets), read aloud by one speaker
Linux-Admin-Sentences	danielrosehill	open	Linux-Admin-Sentence-Bank	Dictation-style English sentences about Linux desktop and server administration (KDE Plasma, GNOME, systemd, storage, networking, homelab), read aloud by one speaker
Personal-Vocabulary-Sentences	danielrosehill	open	Personal-Vocabulary-Sentence-Bank	English sentences built around names a speech model gets wrong for one speaker: his own name, his company and his project names (Daniel Rosehill, DSR Holdings, Carrot Cake AI, the podcast My Weird Prompts and its hosts Corn and Herman Poppl
Read-Aloud-Voice-Sentences	danielrosehill	gated		Every take from the capture app; all speech models train on a pinned revision of it, selected by blend. Gated.
Smart-Home-Sentences	danielrosehill	open	Smart-Home-Sentence-Bank	Dictation-style English sentences about home automation (Home Assistant, Zigbee, Z-Wave, Matter, Thread, ESPHome, MQTT and device brands), read aloud by one speaker
Tech-Sentences	danielrosehill	open	AI-Tech-Vocabulary-Sentence-Bank	Dictation-style English sentences dense in software, DevOps, cloud, self-hosting and AI vocabulary (product names, CLI tools, acronyms, model names), read aloud by one speaker
Tech-Sentences-For-ASR-Training	danielrosehill	open		Early tech sentences with audio; a prompt source for the bank.
Voice-Command-Sentences	danielrosehill	open		One speaker reading desktop voice commands for a KDE Plasma voice-command agent: launching, switching to and closing apps (Konsole, Kate, Dolphin, Kdenlive, Audacity), browser and terminal tabs, windows and virtual desktops ("Move this to d

Benchmarks (2)

Dataset	Account	Access	Pairs with	What it is
English-Hebrew-Codeswitching-ASR-Benchmark	Hebrish	gated		Held-out test sets (v1, v2) the models are scored on; never trained on.
Hebrish-Validation	Hebrish	open		Validation benchmark for the English-Hebrew code-switching speech models.

Training mixes (2)

Dataset	Account	Access	Pairs with	What it is
Parakeet-V2-Blend-All	Hebrish	gated		Export of the `all` training mix, for reproducing Sabich runs.
Parakeet-V2-Blend-No-Hebrew	Hebrish	gated		The exact data of training run r7-general, Daniel's English model without code-switching (not a Hebrish model).

Error lists (2)

Dataset	Account	Access	Pairs with	What it is
ASR-Error-List	Hebrish	gated		Words the models mishear in Daniel's own dictation, with the wrong forms; drives the capture app's priority words. Sabich v2.1 repeated its takes 3 times.
ASR-Error-Prediction	Hebrish	open		Predicted mistranscription candidates for planning new sentences.

Annotations (2)

Dataset	Account	Access	Pairs with	What it is
Room-Tone-Annotations	Hebrish	open		Room-tone recordings and labels for noise conditions.
Take-Labels	Hebrish	open		Hand labels for takes in the master dataset.

Reference data (6)

Dataset	Account	Access	Pairs with	What it is
Background-Noise	Hebrish	gated		Background noise recordings.
Example-Audio-Sentences	Hebrish	gated	Example-Sentence-Bank	Format reference for audio datasets: the project's standard columns, with a schema file and three real takes with the place block blanked.
Example-Sentence-Bank	Hebrish	gated	Example-Audio-Sentences	Format reference for sentence datasets: the project's standard columns, with a schema file and a handful of invented neutral sentences.
Hebrew-Term-Pronunciations-IPA	Hebrish	gated		How Hebrew words are pronounced inside English speech, in IPA, measured from real recordings.
Israeli-Street-Names	Hebrish	gated		Israel's national street list, source for the street-name prompts.
Speakers	Hebrish	open		Accreditation: one row per contributor whose voice is in the project.

Other (1)

Dataset	Account	Access	Pairs with	What it is
Hebrish-Ideas	Hebrish	open		Ideas recorded in the capture app.

Related, outside the Hebrish speech models (7)

Dataset	Account	Access	What it is
Annotated-Voice-Notes	danielrosehill	open	Real voice notes with annotations.
ASR-Cleanup-Before-After	danielrosehill	open	ASR cleanup before and after pairs. Separate project.
ASR-Paragraph-Breaks	danielrosehill	open	Hand-annotated paragraph breaks for flat speech-to-text output. Separate project.
Proofed-ASR-Cleanup-Training-Data	danielrosehill	open	Audio paired with clean, proofed text, for ASR cleanup training. Separate project.
Voice-Note-Audio	danielrosehill	open	Real-world voice recordings with transcripts and annotations.
Voice-Notes	danielrosehill	open	Real voice notes.
Wake-Words	danielrosehill	open	Jewish and Israeli first-name wake words, for local wake word models.

Unclassified (add to data/dataset-index-extra.json) (2)

Dataset	Account	Access	Pairs with	What it is
ASR-Formatter-Training-Pairs	Hebrish	open
Model-Iterations	Hebrish	open

Planned (2)

Registered tasks whose subset appears once takes are recorded: Name-Cipher-Sentences, October-7-Memorial-Sentences.

Models

Model	Task	Data (blend @ revision, or dataset)	Source
Jachnun PoC-1	speech-recognition	`cs` @ b31aeabb	confirmed
Jachnun PoC-2	speech-recognition	`cs` @ 3bdaa2a5	confirmed
Jachnun PoC-3	speech-recognition	`cs` @ b4731062	confirmed
Jachnun PoC-4	speech-recognition	`cs` @ 4fcca820	confirmed
Jachnun PoC-5	speech-recognition	`cs` @ fcf5aec1	confirmed
Shakshuka PoC-1	speech-recognition	`cs` @ fcf5aec1	confirmed
Jachnun v1.0	speech-recognition	`cs` @ 550bfcaa	confirmed
Shakshuka PoC-2	speech-recognition	`cs` @ 550bfcaa	confirmed
Sabich v1.0	speech-recognition	`all` @ f956dbb1	confirmed
Shakshuka v1.0	speech-recognition	`all` @ e918e381	confirmed
Sabich-Parakeet-110M v1.0	speech-recognition	`all` @ e918e381	confirmed
Sabich-Qwen3-ASR-1.7B PoC	speech-recognition	`all` @ e918e381	confirmed
Sabich-Canary-Qwen-2.5B v1.0	speech-recognition	`all` @ cf84f7e4	confirmed
Sabich-asr4all-M v1.0	speech-recognition	`all` @ cf84f7e4	confirmed
Sabich-asr4all-S v1.0	speech-recognition	`all` @ cf84f7e4	confirmed
Sabich v2.0	speech-recognition	`all` @ 110d144a	confirmed
Shakshuka v2.0	speech-recognition	`all` @ 110d144a	confirmed
Sabich v2.1	speech-recognition	`all` @ 2cec2ebc	confirmed
Sabich v2.2	speech-recognition	`all` @ 051646b5	confirmed
Sabich v2.3	speech-recognition	`all` @ ca08e06b	confirmed
Sabich v2.4	speech-recognition	`all` @ 60d0dfa5	confirmed
Shakshuka v2.1	speech-recognition	`all` @ e41edcab	confirmed
Meorav v2.1	text-to-speech	`Hebrish/Hebrish-TTS-Voice-Recordings`	confirmed
Meorav v2.0	text-to-speech	`Hebrish/Hebrish-TTS-Voice-Recordings`	confirmed
Meorav PoC-1	text-to-speech	`Hebrish/Hebrish-TTS-Voice-Recordings`	confirmed
Whisper-Hebrish v2	speech-recognition	`danielrosehill/Read-Aloud-Voice-Sentences`	confirmed
Whisper-Hebrish small, ACFT	speech-recognition	`danielrosehill/Read-Aloud-Voice-Sentences`	confirmed
Token classifier v1	token-classification	`Hebrish/hebrew-latin-code-switching`	confirmed

Every recording is one speaker's voice (Daniel Rosehill). He sometimes changes his speaking speed on purpose and sometimes says a word the way people commonly mispronounce it; the transcript is always the sentence as written.

Downloads last month: 233

Number of rows:

Total file size:

126 kB