fix audio batch inference

modeling_minicpmo.py

@@ -589,7 +589,7 @@ class MiniCPMO(MiniCPMOPreTrainedModel):
                         audio_start_pos = 0
                         for bound in audio_bounds[i]:
                             audio_len = bound[1] - bound[0]
-                            input_embeddings[0, bound[0] : bound[1]] = audio_embs[
+                            input_embeddings[i, bound[0] : bound[1]] = audio_embs[
                                 audio_start_pos : audio_start_pos + audio_len, :
                             ]
                             audio_start_pos += audio_len