|
--- |
|
license: apache-2.0 |
|
datasets: |
|
- Codec-SUPERB/Voxceleb1_test_original |
|
language: |
|
- en |
|
base_model: |
|
- facebook/wav2vec2-base |
|
model-index: |
|
- name: Speaker_Recognition_with_fine_tuned_Wav2Vec2 |
|
results: |
|
- task: |
|
type: audio-classification |
|
name: Audio Classification |
|
dataset: |
|
name: Voxceleb1 |
|
type: Codec-SUPERB/Voxceleb1_test_original |
|
metrics: |
|
- name: Accuracy |
|
type: accuracy |
|
value: 0.95528 |
|
- name: Precision |
|
type: precision |
|
value: 0.937891 |
|
- name: Recall |
|
type: recall |
|
value: 0.944353 |
|
- name: F1 |
|
type: f1 |
|
value: 0.938486 |
|
- name: EER |
|
type: eer |
|
value: 0.013731 |
|
pipeline_tag: audio-classification |
|
library_name: transformers |
|
tags: |
|
- audio-classification |
|
- speech-recognition |
|
- wav2vec2 |
|
- english |
|
metrics: |
|
- accuracy |
|
- precision |
|
- f1 |
|
--- |
|
# Speaker_Recognition_with_fine_tuned_Wav2Vec2 |
|
|
|
### Downstream Use |
|
|
|
Speaker Recognition on the Voxceleb1['test'] split |
|
|
|
## Training Details |
|
|
|
### Training Data |
|
|
|
Training is carried out on [Voxceleb1](https://huggingface.co/datasets/Codec-SUPERB/Voxceleb1_test_original) |
|
|
|
|
|
## Evaluation |
|
|
|
| Metric | Value | |
|
|:---------:|:--------:| |
|
| Accuracy | 0.985641 | |
|
| Precision | 0.984270 | |
|
| Recall | 0.983071 | |
|
| F1-score | 0.983118 | |
|
| EER | 0.003487 | |
|
|
|
### Testing Data, Factors & Metrics |
|
|
|
#### Testing Data |
|
|
|
| Metric | Value | |
|
|:---------:|:--------:| |
|
| Accuracy | 0.95528 | |
|
| Precision | 0.937891 | |
|
| Recall | 0.944353 | |
|
| F1-score | 0.938486 | |
|
| EER | 0.013731 | |