Update app.py

app.py

@@ -11,6 +11,8 @@ model = AutoModelForImageTextToText.from_pretrained(
     device_map="auto"
 )
 
+
+
 def convert_audio_to_text(audio_file):
     # Load audio
     waveform, sample_rate = torchaudio.load(audio_file)
@@ -19,10 +21,18 @@ def convert_audio_to_text(audio_file):
     if waveform.shape[0] > 1:
         waveform = waveform.mean(dim=0, keepdim=True)
 
-    prompt = "Transcribe the audio."
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                {"type": "audio", "audio": audio_file},
+                {"type": "text", "text": "Transcribe this audio into English, and then translate it into French."},
+            ]
+        }
+    ]
 
     inputs = processor.apply_chat_template(
-        prompt,
+        messages,
         add_generation_prompt=True,
         tokenize=True, return_dict=True,
         return_tensors="pt",