File size: 913 Bytes
7837b8e 93b926f 7837b8e 93b926f 7837b8e afc6339 93b926f afc6339 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 |
# This speech tagger performs transcription, annotates entities, predict intent for SLURP dataset
Model is suitable for voiceAI applications.
## Model Details
- **Model type**: NeMo ASR
- **Architecture**: Conformer CTC
- **Language**: English
- **Training data**: Slurp dataset
- **Performance metrics**: [Metrics]
## Usage
To use this model, you need to install the NeMo library:
```bash
pip install nemo_toolkit
```
### How to run
```python
import nemo.collections.asr as nemo_asr
# Step 1: Load the ASR model from Hugging Face
model_name = 'WhissleAI/speech-tagger_en_slurp-iot'
asr_model = nemo_asr.models.EncDecCTCModel.from_pretrained(model_name)
# Step 2: Provide the path to your audio file
audio_file_path = '/path/to/your/audio_file.wav'
# Step 3: Transcribe the audio
transcription = asr_model.transcribe(paths2audio_files=[audio_file_path])
print(f'Transcription: {transcription[0]}')
``` |