| language: | |
| - am | |
| license: mit | |
| tags: | |
| - automatic-speech-recognition | |
| - speech | |
| metrics: | |
| - wer | |
| - cer | |
| # Amharic ASR using fine-tuned Wav2vec2 XLSR-53 | |
| This is a finetuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) trained on the [Amharic Speech Corpus](http://www.openslr.org/25/). This corpus was produced by [Abate et al. (2005)](https://www.isca-speech.org/archive/interspeech_2005/abate05_interspeech.html) (10.21437/Interspeech.2005-467). | |
| ## Usage | |
| The model can be used as follows: | |
| ```python | |
| import librosa | |
| from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor | |
| model = Wav2Vec2ForCTC.from_pretrained("agkphysics/wav2vec2-large-xlsr-53-amharic") | |
| processor = Wav2Vec2Processor.from_pretrained("agkphysics/wav2vec2-large-xlsr-53-amharic") | |
| audio, _ = librosa.load("/path/to/audio.wav", sr=16000) | |
| input_values = processor( | |
| audio.squeeze(), | |
| sampling_rate=16000, | |
| return_tensors="pt" | |
| ).input_values | |
| model.eval() | |
| with torch.no_grad(): | |
| logits = model(input_values).logits | |
| preds = logits.argmax(-1) | |
| texts = processor.batch_decode(preds) | |
| print(texts[0]) | |
| ``` | |
| ## Training | |
| The code to train this model is available at https://github.com/agkphysics/amharic-asr. | |