Feature Extraction
MLX
resnet34_embedding
speaker-recognition
speaker-embedding
speaker-diarization
audio
resnet
apple-silicon
Instructions to use mlx-community/wespeaker-voxceleb-resnet34-LM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/wespeaker-voxceleb-resnet34-LM with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir wespeaker-voxceleb-resnet34-LM mlx-community/wespeaker-voxceleb-resnet34-LM
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
Upload config.json with huggingface_hub
Browse files- config.json +23 -0
config.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "resnet34_embedding",
|
| 3 |
+
"architecture": "ResNet34",
|
| 4 |
+
"task": "speaker-embedding",
|
| 5 |
+
"framework": "mlx",
|
| 6 |
+
"feat_dim": 80,
|
| 7 |
+
"embed_dim": 256,
|
| 8 |
+
"m_channels": 32,
|
| 9 |
+
"pooling": "TSTP",
|
| 10 |
+
"num_blocks": [3, 4, 6, 3],
|
| 11 |
+
"dtype": "float32",
|
| 12 |
+
"sample_rate": 16000,
|
| 13 |
+
"num_parameters": 6600000,
|
| 14 |
+
"model_size_mb": 25,
|
| 15 |
+
"source": "pyannote/wespeaker-voxceleb-resnet34-LM",
|
| 16 |
+
"conversion_date": "2026-01-16",
|
| 17 |
+
"validation": {
|
| 18 |
+
"speaker_similarity_max_diff": 0.024,
|
| 19 |
+
"speaker_similarity_mean_diff": 0.008,
|
| 20 |
+
"numerical_max_abs_diff": 0.17,
|
| 21 |
+
"status": "validated"
|
| 22 |
+
}
|
| 23 |
+
}
|