Translation
NeMo
French
asr
speech
automatic-speech-recognition
canary-v2
french-media
tv-transcription
Eval Results (legacy)
Instructions to use Archime/canary-v2-fr-tv-media with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use Archime/canary-v2-fr-tv-media with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("Archime/canary-v2-fr-tv-media") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
Upadate README.md
Browse files
README.md
CHANGED
|
@@ -150,14 +150,22 @@ Canary-v2 requires specific input prompts to define the task (ASR/Translation) a
|
|
| 150 |
|
| 151 |
### Installation
|
| 152 |
```bash
|
| 153 |
-
pip install nemo_toolkit['all']==2.6.
|
| 154 |
```
|
| 155 |
|
| 156 |
```python
|
| 157 |
import nemo.collections.asr as nemo_asr
|
|
|
|
| 158 |
|
| 159 |
# Load the model
|
| 160 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 161 |
|
| 162 |
# Transcription with specific prompts
|
| 163 |
# pnc="yes" enables punctuation and capitalization
|
|
|
|
| 150 |
|
| 151 |
### Installation
|
| 152 |
```bash
|
| 153 |
+
pip install nemo_toolkit['all']==2.6.0
|
| 154 |
```
|
| 155 |
|
| 156 |
```python
|
| 157 |
import nemo.collections.asr as nemo_asr
|
| 158 |
+
from huggingface_hub import hf_hub_download
|
| 159 |
|
| 160 |
# Load the model
|
| 161 |
+
REPO_ID="Archime/canary-v2-fr-tv-media"
|
| 162 |
+
MODEL_NAME="canary_v2_fr_tv_media.nemo"
|
| 163 |
+
|
| 164 |
+
model_path = hf_hub_download(
|
| 165 |
+
repo_id=REPO_ID,
|
| 166 |
+
filename=MODEL_NAME
|
| 167 |
+
)
|
| 168 |
+
model = nemo_asr.models.EncDecMultiTaskModel.restore_from(model_path)
|
| 169 |
|
| 170 |
# Transcription with specific prompts
|
| 171 |
# pnc="yes" enables punctuation and capitalization
|