Update app.py

app.py

@@ -6,28 +6,32 @@ from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
 import torch
 import torchaudio
 
+# ✅ Explicitly set the backend so .wav files load properly
+torchaudio.set_audio_backend("soundfile")
+
 app = FastAPI()
 
+# ✅ Load model and processor
 processor = Wav2Vec2Processor.from_pretrained("tacab/ASR_SOMALI")
 model = Wav2Vec2ForCTC.from_pretrained("tacab/ASR_SOMALI")
 model.to("cpu")
 
 @app.post("/transcribe")
 async def transcribe(file: UploadFile = File(...)):
+    # ✅ Save uploaded file to /tmp
     audio_bytes = await file.read()
-
     temp_path = "/tmp/temp.wav"
     with open(temp_path, "wb") as f:
         f.write(audio_bytes)
 
+    # ✅ Load audio file
     speech_array, sampling_rate = torchaudio.load(temp_path)
-    inputs = processor(speech_array.squeeze(), return_tensors="pt", sampling_rate=sampling_rate)
 
+    # ✅ Run through ASR model
+    inputs = processor(speech_array.squeeze(), return_tensors="pt", sampling_rate=sampling_rate)
     with torch.no_grad():
         logits = model(**inputs).logits
-
     predicted_ids = torch.argmax(logits, dim=-1)
     transcription = processor.batch_decode(predicted_ids)[0]
 
     return {"text": transcription}
-