[CI][UT]enable multimodal ut

pytest.ini

@@ -16,7 +16,6 @@ norecursedirs =
     vllm-empty/tests/lora
     vllm-empty/tests/models
     vllm-empty/tests/mistral_tool_use
-    vllm-empty/tests/multimodal
     vllm-empty/tests/standalone_tests
     vllm-empty/tests/async_engine
     vllm-empty/tests/mq_llm_engine
@@ -51,6 +50,9 @@ addopts = --ignore=vllm-empty/tests/test_utils.py
           --ignore=vllm-empty/tests/models/decoder_only/language/test_aqlm.py
           --ignore=vllm-empty/tests/models/decoder_only/language/test_gptq_marlin.py
           --ignore=vllm-empty/tests/models/decoder_only/language/test_gptq_marlin_24.py
+          --ignore=vllm-empty/tests/multimodal/test_processing.py
+          --ignore=vllm-empty/tests/multimodal/test_processor_kwargs.py
+          --ignore=vllm-empty/tests/multimodal/test_utils.py
 
 testpaths =
     vllm-empty/tests

vllm_ascend/model_runner.py

@@ -709,6 +709,7 @@ def _compute_multi_modal_input(self, inter_data: InterDataForSeqGroup,
             assert image_grid_thw is not None or video_grid_thw is not None, (
                 "mrope embedding type requires multi-modal input mapper "
                 "returns 'image_grid_thw' or 'video_grid_thw'.")
+            second_per_grid_ts = mm_kwargs.get("second_per_grid_ts", None)
 
             hf_config = self.runner.model_config.hf_config
 
@@ -721,14 +722,10 @@ def _compute_multi_modal_input(self, inter_data: InterDataForSeqGroup,
                 mrope_input_positions, mrope_position_delta = \
                     MRotaryEmbedding.get_input_positions(
                         token_ids,
+                        hf_config,
                         image_grid_thw=image_grid_thw,
                         video_grid_thw=video_grid_thw,
-                        image_token_id=hf_config.image_token_id,
-                        video_token_id=hf_config.video_token_id,
-                        vision_start_token_id=hf_config.vision_start_token_id,
-                        vision_end_token_id=hf_config.vision_end_token_id,
-                        spatial_merge_size=hf_config.vision_config.
-                        spatial_merge_size,
+                        second_per_grid_ts=second_per_grid_ts,
                         context_len=inter_data.context_lens[seq_idx],
                         seq_len=inter_data.seq_lens[seq_idx],
                     )