Update app.py

app.py

@@ -14,11 +14,16 @@ device = "cuda:0" if torch.cuda.is_available() else "cpu"
 
 # load model and processor
 model_id = "rbcurzon/whisper-small-fil"
-pipe = pipeline("automatic-speech-recognition", 
-                model=model_id,
-                chunk_length_s=30,
-                device=device)
+
+pipe = pipeline(
+    "automatic-speech-recognition", 
+    model=model_id,
+    chunk_length_s=30,
+    device=device
+)
+
 model = load_silero_vad()
+
 client = genai.Client(api_key=os.environ.get("GENAI_API_KEY")) # Do not share api key
 
 """**FastAPI**"""
@@ -58,11 +63,17 @@ async def test(
             sampling_rate=16000
         )
 
+        generate_kwargs = {
+            "language": "tagalog",
+            "return_timestamps": True,
+            "condition_on_prev_tokens": False,
+        }
+        
         result = pipe(
             "only_speech.wav", # Transcribe audio
             batch_size=8, 
             return_timestamps=True,
-            generate_kwargs={"language": "tagalog","return_timestamps": True,}
+            generate_kwargs=generate_kwargs
         )
         print(result)
 
@@ -103,4 +114,11 @@ async def test(text: str,
     result = translate(text, srcLang, tgtLang)
     print('Raw: ', text)
     print('Translated: ', result)
-    return {'translated_text': result}
+    
+    result_dict = {
+        "text": text, 
+        "translated_text": result, 
+        "srcLang": srcLang, 
+        "tgtLang": tgtLang
+    }
+    return result_dict