| import gradio as gr | |
| import torch | |
| import torchaudio | |
| from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor | |
| import librosa | |
| import numpy as np | |
| import re | |
| device = torch.device("cuda" if torch.cuda.is_available() else "cpu") | |
| processor = Wav2Vec2Processor.from_pretrained("the-cramer-project/Wav2vec-Kyrgyz") | |
| model = Wav2Vec2ForCTC.from_pretrained("the-cramer-project/Wav2vec-Kyrgyz") | |
| model.to(device = device) | |
| def transcribe(file_): | |
| arr_audio, _ = librosa.load(file_, sr=16000) | |
| inputs = processor(arr_audio, sampling_rate=16_000, return_tensors="pt", padding=True) | |
| with torch.no_grad(): | |
| logits = model(inputs.input_values.to(device = device), attention_mask=inputs.attention_mask.to(device = device)).logits | |
| pred_ids = torch.argmax(logits, dim=-1) | |
| text = processor.batch_decode(pred_ids)[0] | |
| return text | |
| iface = gr.Interface( | |
| fn=transcribe, | |
| inputs=gr.Audio(type="filepath"), | |
| outputs="text", | |
| title="Wave2Vec Kyrgyz", | |
| description="Realtime demo for Kyrgyz speech recognition using a wave2vec model.", | |
| ) | |
| iface.launch() |