You need to agree to share your Hugging Face information to access this model
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
This model is outdated and no longer updated. For most people Hebrish/Sabich-Rolling is the right choice; Hebrish/Falafel for voice commands; Fritter-ASR/Fritter for English without Hebrew. Guide: https://huggingface.co/spaces/Hebrish/Index. Access is approved automatically once you accept the terms. Use is governed by the Hebrish terms of use (https://hebrish.app/terms): non-commercial use is free; commercial use needs a commercial licence from access@hebrish.app.
Log in or Sign Up to review the conditions and access this model content.
Sabich v2.3: ONNX fp32 and int8
Outdated: this model is no longer updated. It stays available for reference, reproducibility and existing installs. The models kept up to date are:
- Sabich Rolling: the one for most people. English with Hebrew words mixed in, plus tech vocabulary; retrained as new recordings come in, and what the Hebrish apps run.
- Falafel: voice commands, the kind of thing you say to an AI agent or your computer.
- Fritter: general English and tech vocabulary without Hebrew code-switching (a separate project, CC-BY-NC-4.0).
Not sure which to pick? See the Hebrish Index or https://hebrish.app/models/choose.
ONNX export of Sabich v2.3 (2026-10-30), the kitchen-sink Parakeet TDT 0.6B v2
fine-tune: English-Hebrew code-switching plus tech vocabulary, brand names, numbers in digits and Irish-English
phrasing, in one speaker's voice. Read the main card for results and limitations. Earlier versions of this repo held Sabich v1.0 (run r7-all); they stay in the commit history.
Try it in your browser: Sabich Demo runs the int8 build of these files on your own device.
Website: hebrish.app. Use is governed by the Hebrish terms of use: non-commercial use is free, commercial use needs a licence from access@hebrish.app.
The layout is the one used by onnx-asr and
istupakov/parakeet-tdt-0.6b-v2-onnx, exported the
same way (NeMo model.export) from training run r13-all (Hebrish/Sabich at c818843b), the same files layout as
Hebrish/Jachnun-ONNX:
| File | Precision |
|---|---|
encoder-model.onnx + encoder-model.onnx.data (2.4 GB) |
fp32 |
decoder_joint-model.onnx |
fp32 |
encoder-model.int8.onnx (652 MB) |
int8, dynamic (QUInt8) |
decoder_joint-model.int8.onnx (9 MB) |
int8, dynamic (QInt8) |
vocab.txt, config.json |
nemo-conformer-tdt, 128 mel features |
Exported 2026-10-30 by the same pipeline as v2.2, whose fp32 and int8 builds produced identical text on five held-out phone clips and matched the sherpa-onnx int8 and GGUF F16 builds. Results for v2.3 are on the main model card.
onnx-asr
# pip install onnx-asr[cpu] huggingface_hub
from huggingface_hub import snapshot_download
import onnx_asr
path = snapshot_download("Hebrish/Sabich-ONNX",
allow_patterns=["*.int8.onnx", "vocab.txt", "config.json"])
model = onnx_asr.load_model("nemo-conformer-tdt", path, quantization="int8")
print(model.recognize("audio.wav")) # 16 kHz mono
Drop allow_patterns and quantization for fp32. Voxtype loads this
layout too; give it a directory holding one precision only (instructions on the
Jachnun-ONNX card).
Base model and licence notices
This model is a build of Hebrish/Sabich, a fine-tune of nvidia/parakeet-tdt-0.6b-v2, by NVIDIA, licensed under CC-BY-4.0. Hebrish modified the base model by fine-tuning it and converted the weights for this build.
The licence stated at the top of this card covers Hebrish's contribution; the base model's licence keeps applying to the base model, and nothing here restricts what that licence allows for it. Terms of use: https://hebrish.app/terms.
Hebrish is not affiliated with or endorsed by NVIDIA.
- Downloads last month
- 7
