Automatic Speech Recognition
NeMo
Finnish
asr
speech-recognition
canary-v2
kenlm
finnish
Eval Results (legacy)
Instructions to use RASMUS/Finnish-ASR-Canary-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use RASMUS/Finnish-ASR-Canary-v2 with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("RASMUS/Finnish-ASR-Canary-v2") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
Speaker Tasks Dataset Scripts
In this folder are scripts to download speaker tasks (mainly for diarization) datasets. These scripts will return NeMo format manifest files to use with Diarization.
We also have scripts for CallHome and DIHARD3, however the data has to be downloaded separately. If you require the scripts please leave an issue.