DF Arena 100M - Antispoofing Model
We are excited to release DF Arena 100M Antispoofing model trained on large scale speech antispoofing datasets. Check out the release on DF Arena leaderboard
Live Demo
Checkout our Huggingface Space for interactive demo of the model
Evaluation
| Dataset | EER (%) | F1-score | Accuracy (%) |
|---|---|---|---|
| dfadd | 0.00 | 0.9993 | 99.97 |
| add_2023_round_2 | 17.04 | 0.8779 | 82.96 |
| codecfake | 8.75 | 0.8639 | 91.25 |
| asvspoof_2021_la | 7.61 | 0.7083 | 92.39 |
| in_the_wild | 2.22 | 0.9822 | 97.78 |
| asvspoof_2019 | 1.54 | 0.9299 | 98.47 |
| add_2022_track_1 | 27.06 | 0.6073 | 72.94 |
| fake_or_real | 7.42 | 0.9260 | 92.60 |
| asvspoof_2024 | 21.40 | 0.5995 | 78.60 |
| add_2022_track_3 | 5.42 | 0.8510 | 94.58 |
| add_2023_round_1 | 8.90 | 0.9360 | 91.10 |
| librisevoc | 0.22 | 0.9927 | 99.79 |
| asvspoof_2021_df | 5.50 | 0.4889 | 94.50 |
| sonar | 4.48 | 0.9611 | 95.54 |
| Average | 8.40 | 0.8374 | 91.60 |
| Pooled | 13.92 | 0.7378 | 86.08 |
License and Usage
The model uses a closed license. Kindly contact ajinkya.kulkarni@idiap.ch for access to the model.