Create app.py

app.py

@@ -0,0 +1,31 @@
+import gradio as gr
+from transformers import pipeline
+import numpy as np
+import librosa
+from punctuators.models import PunctCapSegModelONNX
+
+transcriber = pipeline("automatic-speech-recognition", model="Oysiyl/w2v-bert-2.0-dutch-colab-CV16.0")
+punct_cap_model = PunctCapSegModelONNX.from_pretrained("1-800-BAD-CODE/xlm-roberta_punctuation_fullstop_truecase")
+
+def transcribe(audio):
+    sr, y = audio
+    y = y.astype(np.float32)
+    y /= np.max(np.abs(y))
+    if sr != 16000:
+        y = librosa.resample(y, orig_sr=sr, target_sr=16000)
+    transcribed_text = transcriber({"sampling_rate": 16000, "raw": y})["text"]
+    punct_cap_text = punct_cap_model.infer(texts=[transcribed_text], apply_sbd=True)[0][0]
+    return punct_cap_text
+
+
+demo = gr.Interface(
+    transcribe,
+    gr.Audio(sources=["upload", "microphone"]),
+    outputs="text",
+    title="Automatic Speech Recognition for Dutch language demo",
+    description="Click on the example below, upload audio from file or say something in microphone!",
+    examples=[["examples/example1.wav"], ["examples/example2.wav"]],
+    cache_examples=True
+)
+
+demo.launch()