Downloads · 30 days
0
parakeet-primeline-sherpa-onnx-int8
Joni000000000/parakeet-primeline-sherpa-onnx-int8
parakeet-primeline-sherpa-onnx-int8 is a automatic speech recognition model from Joni000000000. Use it when you need speech turned into text. The card lists the license as cc-by-4.0.
A sherpa-onnx-compatible ONNX export of primeline/parakeet-primeline, which is itself a German finetune of nvidia/parakeet-tdt-0.6b-v3.
Access
Public
Updated Sep 9, 2026
Repo size
973 MB
Likes
2
Public
Repo makeup
Click a slice to open those files.
.bz2487 MB · 100%
At a glance
- Task
- Automatic Speech Recognition
- License
- cc-by-4.0
- Access
- Public
- Created
- Sep 8, 2026
- Updated
- Sep 9, 2026
- SHA
- 182bf371
Model card
From the Hugging Face model README
parakeet-primeline, sherpa-onnx export (int8)
A sherpa-onnx-compatible ONNX export of primeline/parakeet-primeline, which is itself a German finetune of nvidia/parakeet-tdt-0.6b-v3.
Produced with sherpa-onnx's own first-party export script
(scripts/nemo/parakeet-tdt-0.6b-v3/export_onnx.py), unmodified except for the
provenance strings in the ONNX metadata, which upstream hardcodes to nvidia's
URL. The exporting wrapper is scripts/export-primeline-onnx.sh in the yappr
repository.
Contents: encoder.int8.onnx, decoder.int8.onnx, joiner.int8.onnx,
tokens.txt (and bpe.vocab, for sherpa's hotword support). Offline
transducer; load it with sherpa-onnx's OfflineRecognizer and
model_type = "nemo_transducer".
Licensed CC-BY-4.0, inherited from the upstream model. Original work by primeline; this repository redistributes a format conversion of it.
- Task
- Automatic Speech Recognition
- License
- cc-by-4.0
- Languages
- de
- Created
- Sep 8, 2026
- Updated
- Sep 9, 2026