Automatic Speech Recognition
ONNX
NeMo
Lithuanian
nemo-conformer-rnnt
nemotron
rnnt
onnx-asr
onnxruntime
lithuanian
speech-recognition
wyoming
home-assistant
Eval Results (legacy)
Instructions to use RobertasTa/lt-ausys with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use RobertasTa/lt-ausys with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("RobertasTa/lt-ausys") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
LT_ausys: model files are now datolabs/upelis-0.6b-lt-v1 in ONNX (Lithuanian prompt built in, fp32 + int8 MatMul-only), OpenMDW-1.1 LICENSE + NOTICE, new card. Previous kmynas files: revision v1-kmynas
4b1cfb9 verified Download encoder-model.onnx.data from RobertasTa/lt-ausys: direct link, hf CLI and curl.
- Browser
- Download file 2.45 GB
-
https://huggingface.co/RobertasTa/lt-ausys/resolve/main/encoder-model.onnx.data
- Command line
-
hf download hf://RobertasTa/lt-ausys/encoder-model.onnx.data
-
curl -L -o encoder-model.onnx.data https://huggingface.co/RobertasTa/lt-ausys/resolve/main/encoder-model.onnx.data
2.45 GB
- Xet hash:
- 67a4e8d212298c3dc19b09874dbc1cb8d905455b3507ac6a1f4ff54e48f0a8c8
- Size of remote file:
- 2.45 GB
- SHA256:
- 006af3325702f0e944987c8e4fd365d0845666993e0450963416fac45ad55783
·
Xet efficiently stores Large Files inside Git, intelligently splitting files into unique chunks and accelerating uploads and downloads. More info.