filipino-wav2vec2-l-xls-r-300m-official
Khalsuu/filipino-wav2vec2-l-xls-r-300m-official
A popular open speech-to-text model, with 2.3M downloads a month. gigarouter benchmarks and hosts it as an OpenAI-compatible API.
about this model
filipino-wav2vec2-l-xls-r-300m-official is an automatic speech recognition (ASR) model fine-tuned from facebook/wav2vec2-xls-r-300m on the filipino_voice dataset. It is designed for transcribing Filipino speech to text.
Key strengths
Fine-tuned specifically for the Filipino language, leveraging the large XLS-R 300M pretrained model. The model achieves a word error rate (WER) of 0.2922 on the evaluation set after 30 epochs of training.
Benchmark results
| Metric | Value |
|---|---|
| Validation Loss | 0.4672 |
| Word Error Rate (WER) | 0.2922 |
The model was trained with a learning rate of 0.0003, batch size 8, gradient accumulation steps 2, and mixed precision training.
We're benchmarking and onboarding filipino-wav2vec2-l-xls-r-300m-official as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.