Sabich trained forward from its previous weights on all recordings so far. The Hebrish apps default since 2026-11-02.
AI & ML interests
Improving voice AI accuracy for code-switched English-Hebrew for English-speaking immigrants in Israel and the Jewish world. Models: ASR (focus), TTS ( POC / experimentary).
Recent Activity
View all activity
Third-party models the Hebrish apps download and run on the device, alongside the speech model.
ASR model variants for compute constraint hardware, esp. wearables. WER is not yet production-appropriate
Try the models, browse the data and the benchmark. hebrish.app
-
English-Hebrew Code-Switching ASR Leaderboard
🎙ASR on English speech with Hebrew words in it
-
Hebrew-in-Latin Span Review
🔤Review and label Hebrew words in Latin script
-
English-Hebrew Code-Switching Browser
🎙Listen to English-Hebrew code-switching by term
-
English-Hebrew Code-Switching Parakeet
🦜Versions, downloads, results and notes for the fine-tune
Text only: the sentence banks read aloud to record the training data. Recordings: 'Hebrish: audio datasets'. hebrish.app
Sabich: Hebrew vocabulary + tech vocabulary. Code-switching plus tech terms, brand names and numbers. hebrish.app
Every Hebrish model, dataset and Space, including those still on danielrosehill. hebrish.app
-
danielrosehill/Whisper-Hebrish
Automatic Speech Recognition • 0.8B • Updated • 43 -
Hebrish/Whisper-Hebrish-Small-ACFT
Automatic Speech Recognition • 0.2B • Updated • 35 -
danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching
Automatic Speech Recognition • 0.6B • Updated • 3.31k -
Hebrish/Jachnun-GGUF
Automatic Speech Recognition • 0.6B • Updated • 222
Sabich v2.3 (2026-10-30): the recommended Hebrish model, every format, its benchmark and demo. hebrish.app/models/sabich
Code-switched English- Hebrew TTS model
Recorded speech, one dataset per phenomenon (code-switching, Yiddish, brand names, numbers, noise …), plus the benchmark. hebrish.app
Jachnun: English + Hebrew vocabulary. The Hebrish model, its builds for every runtime, and demos. hebrish.app
Audio the Parakeet v2 fine-tunes blend from: master dataset, category and task subsets, imports, one blend dataset per model.
Sabich trained forward from its previous weights on all recordings so far. The Hebrish apps default since 2026-11-02.
Third-party models the Hebrish apps download and run on the device, alongside the speech model.
Sabich v2.3 (2026-10-30): the recommended Hebrish model, every format, its benchmark and demo. hebrish.app/models/sabich
ASR model variants for compute constraint hardware, esp. wearables. WER is not yet production-appropriate
Code-switched English- Hebrew TTS model
Try the models, browse the data and the benchmark. hebrish.app
-
English-Hebrew Code-Switching ASR Leaderboard
🎙ASR on English speech with Hebrew words in it
-
Hebrew-in-Latin Span Review
🔤Review and label Hebrew words in Latin script
-
English-Hebrew Code-Switching Browser
🎙Listen to English-Hebrew code-switching by term
-
English-Hebrew Code-Switching Parakeet
🦜Versions, downloads, results and notes for the fine-tune
Recorded speech, one dataset per phenomenon (code-switching, Yiddish, brand names, numbers, noise …), plus the benchmark. hebrish.app
Text only: the sentence banks read aloud to record the training data. Recordings: 'Hebrish: audio datasets'. hebrish.app
Jachnun: English + Hebrew vocabulary. The Hebrish model, its builds for every runtime, and demos. hebrish.app
Sabich: Hebrew vocabulary + tech vocabulary. Code-switching plus tech terms, brand names and numbers. hebrish.app
Audio the Parakeet v2 fine-tunes blend from: master dataset, category and task subsets, imports, one blend dataset per model.
Every Hebrish model, dataset and Space, including those still on danielrosehill. hebrish.app
-
danielrosehill/Whisper-Hebrish
Automatic Speech Recognition • 0.8B • Updated • 43 -
Hebrish/Whisper-Hebrish-Small-ACFT
Automatic Speech Recognition • 0.2B • Updated • 35 -
danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching
Automatic Speech Recognition • 0.6B • Updated • 3.31k -
Hebrish/Jachnun-GGUF
Automatic Speech Recognition • 0.6B • Updated • 222