DSN 2026 language ID baseline

Fine-tuned from facebook/wav2vec2-xls-r-300m on the competition training data only (no external data). Training: fp16 mixed precision, language-balanced sampling, noise and phone-quality augmentation, 1 epochs. Validation language-weighted F1: 0.007.

Downloads last month
-
Safetensors
Model size
0.3B params
Tensor type
F16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support