Remove qwen_omni_utils import and set audio inputs directly

transcript.py

@@ -77,8 +77,7 @@ def main():
     model = Qwen2_5OmniForConditionalGeneration.from_pretrained("Qwen/Qwen2.5-Omni-7B", torch_dtype="auto", device_map="auto", trust_remote_code=True)
     processor = Qwen2_5OmniProcessor.from_pretrained("Qwen/Qwen2.5-Omni-7B", trust_remote_code=True)
 
-    # Import utils after loading model
-    from qwen_omni_utils import process_mm_info
+    # Note: qwen_omni_utils not available, using direct processing
 
     # Load audio
     audio, sr = librosa.load(audio_file, sr=16000)
@@ -107,7 +106,9 @@ def main():
 
         # Preparation for inference
         text = processor.apply_chat_template(conversation, add_generation_prompt=True, tokenize=False)
-        audios, images, videos = process_mm_info(conversation, use_audio_in_video=False)
+        audios = [audio_chunk]
+        images = []
+        videos = []
         inputs = processor(text=text, audio=audios, images=images, videos=videos, return_tensors="pt", padding=True, use_audio_in_video=False)
         inputs = inputs.to(model.device).to(model.dtype)