Create transcribe.py

transcribe.py

@@ -0,0 +1,19 @@
+from pathlib import Path
+from transformers import pipeline
+from config import ASR_MODEL
+
+_asr_pipe = None
+
+def get_asr():
+    global _asr_pipe
+    if _asr_pipe is None:
+        _asr_pipe = pipeline(
+            "automatic-speech-recognition",
+            model=ASR_MODEL
+        )
+    return _asr_pipe
+
+def transcribe(wav_path: Path) -> str:
+    pipe = get_asr()
+    out = pipe(str(wav_path))
+    return out["text"] if isinstance(out, dict) else str(out)