Simplify to audio file input (mic requires PortAudio on NixOS)

transcribe.py

@@ -10,21 +10,22 @@ model = CohereAsrForConditionalGeneration.from_pretrained(
     device_map="auto"
 )
 
-# Download demo audio from Hugging Face
+def transcribe_audio(audio, language="en"):
+    inputs = processor(audio, sampling_rate=16000, return_tensors="pt", language=language)
+    inputs.to(model.device, dtype=model.dtype)
+    
+    outputs = model.generate(**inputs, max_new_tokens=256)
+    text = processor.decode(outputs, skip_special_tokens=True)
+    return text
+
+# Use demo audio file from Hugging Face
+print("Loading demo audio...")
 audio_file = hf_hub_download(
     repo_id="CohereLabs/cohere-transcribe-03-2026",
     filename="demo/voxpopuli_test_en_demo.wav",
 )
-
-# Load audio
 audio = load_audio(audio_file, sampling_rate=16000)
 
-# Process and transcribe
 print("Transcribing...")
-inputs = processor(audio, sampling_rate=16000, return_tensors="pt", language="en")
-inputs.to(model.device, dtype=model.dtype)
-
-outputs = model.generate(**inputs, max_new_tokens=256)
-text = processor.decode(outputs, skip_special_tokens=True)
-
+text = transcribe_audio(audio)
 print(f"\nTranscription:\n{text}\n")