[bugfix] fix multimodal mtp (#21)

Author: Jintao-Huang

src/mcore_bridge/model/gpt_model.py

@@ -399,7 +399,8 @@ def _postprocess(
         output_weight = None
         if self.share_embeddings_and_output_weights:
             output_weight = self.shared_embedding_or_output_weight()
-        if self.config.is_multimodal and self.config.context_parallel_size > 1:
+        if self.config.is_multimodal and self.config.context_parallel_size > 1 and input_ids is not None:
+            # input_ids is required by MTP.
             input_ids = split_cp_inputs(input_ids, getattr(packed_seq_params, 'cu_seqlens_q', None), 1)
 
         if self.mtp_process and labels is not None: