Automatic Speech Recognition
NeMo
Georgian
speech
audio
georgian
telephony
punctuation
FastConformer
Transducer
CTC
hybrid
NeMo
Instructions to use lukab/ka-fastconformer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use lukab/ka-fastconformer with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("lukab/ka-fastconformer") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
Commit History
4 days ago
-
Trained for conversation and phone calls; writes punctuation
Verified signatureFine-tuned further on 429 h with phone-line, VoIP and noise augmentation and punctuated labels. Model card and training config updated.c254baf -
ka-fastconformer-v2: NVIDIA's Georgian FastConformer fine-tuned on 357 h
Verified signature4e8b060 -
initial commit
Verified signatureb2141d0