[Model] Add Ernie4.5 VL Model Support (#22514)

Signed-off-by: wangyafeng <wangyafeng@baidu.com>

examples/offline_inference/vision_language.py

@@ -173,6 +173,37 @@ def run_deepseek_vl2(questions: list[str], modality: str) -> ModelRequestData:
     )
 
 
+# Ernie4.5-VL
+def run_ernie45_vl(questions: list[str], modality: str) -> ModelRequestData:
+    model_name = "baidu/ERNIE-4.5-VL-28B-A3B-PT"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=4096,
+        max_num_seqs=5,
+        limit_mm_per_prompt={modality: 1},
+        trust_remote_code=True,
+    )
+
+    if modality == "image":
+        placeholder = "Picture 1:<|IMAGE_START|><|image@placeholder|><|IMAGE_END|>"
+    elif modality == "video":
+        placeholder = "Video 1:<|VIDEO_START|><|video@placeholder|><|VIDEO_END|>"
+
+    prompts = [
+        (
+            f"<|begin_of_sentence|>User: {question}{placeholder}\n"
+            "Assistant: <think></think>"
+        )
+        for question in questions
+    ]
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
 # Florence2
 def run_florence2(questions: list[str], modality: str) -> ModelRequestData:
     assert modality == "image"
@@ -1602,6 +1633,7 @@ model_example_map = {
     "chameleon": run_chameleon,
     "command_a_vision": run_command_a_vision,
     "deepseek_vl_v2": run_deepseek_vl2,
+    "ernie45_vl": run_ernie45_vl,
     "florence2": run_florence2,
     "fuyu": run_fuyu,
     "gemma3": run_gemma3,