Tag

code-switching

3 posts · all tags

Can speech-to-text spell ‘makolet’? I tested seven models on Hebrish

Can speech-to-text spell ‘makolet’? I tested seven models on Hebrish

English speakers in Israel talk Hebrish, and dictation software mangles the Hebrew words. I built an open benchmark to measure it: my fine-tune gets two-thirds of them right, off-the-shelf models a third to under half.

speech-recognition code-switching hebrew +4
Dictation that finally understands Hebrish, on your phone and offline

Dictation that finally understands Hebrish, on your phone and offline

I fine-tuned NVIDIA's Parakeet v2 speech model to catch the Hebrew words in Israeli English. Release 2 is out: offline downloads for Android and Linux, the full model, and the training data.

speech-recognition parakeet code-switching +4
Teaching a podcast to say makolet: a classifier for Hebrew hiding in English

Teaching a podcast to say makolet: a classifier for Hebrew hiding in English

I released a small model on Hugging Face that finds romanised Hebrew words inside English text, so TTS voices like Chatterbox can stop reading them as English. Built for My Weird Prompts.

hebrew text-to-speech hugging-face +3