Update main.py

main.py

@@ -1,4 +1,6 @@
 import torch
+import torchaudio
+import numpy as np
 from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
 from datasets import load_dataset
 from googletrans import Translator
@@ -45,13 +47,19 @@ async def process_audio(file: UploadFile = File(...)):
             content = await file.read()
             saved_file.write(content)
 
-        # JP
-        original = pipe(file_location)
-        original_version = original["text"]
+        # Load audio file using torchaudio
+        waveform, sample_rate = torchaudio.load(file_location)
+
+        # Ensure the waveform is a 1D array
+        waveform = waveform[0] if waveform.size(0) > 1 else waveform
+
+        # ASR
+        original = pipe(waveform.numpy(), sampling_rate=sample_rate)
+        original_version = original[0]["text"]
 
         # EN
-        result = pipe(file_location, generate_kwargs={"task": "translate"})
-        hasil = result["text"]
+        result = pipe(waveform.numpy(), generate_kwargs={"task": "translate"}, sampling_rate=sample_rate)
+        hasil = result[0]["text"]
 
         # ID
         detect = detect_google(hasil)