Merge branch 'main' into main

inference.py

@@ -480,6 +480,7 @@ class OmniInference:
         with self.fabric.init_tensor():
             model.set_kv_cache(batch_size=2, device=self.device)
 
+
         # Load the audio and process it using Whisper
         mel, leng = load_audio(audio_path)
         audio_feature, input_ids = get_input_ids_whisper_ATBatch(mel, leng, self.whispermodel, self.device)

server.py

@@ -72,6 +72,7 @@ def serve(ip='0.0.0.0', port=60808, device=None):
     OmniChatServer(ip, port=port, run_app=True, device=device)
 
 
+
 if __name__ == "__main__":
     import fire