Armenian ASR

Modern speech-to-text model for Armenian language using Wav2Vec2-BERT (see also with quantization and LoRA fine-tuning).

Setup

pip install -r requirements.txt

Usage

python src/main.py

Model Performance

Best model performance on Armenian Common Voice dataset:

Pre-trained Model: GoshKolotyan/w2v-bert-2.0-armenian-CV16.0-version-demo

Configuration

Edit configs/configs.yml for dataset and training settings.

For memory-efficient training, use configs/quantize_configs.yml.