[Model][VLM] Support R-4B Model (#23246)

Signed-off-by: yannqi <yannqi@qq.com>
Signed-off-by: 杨奇(yann qi) <51905299+yannqi@users.noreply.github.com>
Signed-off-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Co-authored-by: yannqiyang <yannqiyang@tencent.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>

examples/offline_inference/vision_language.py

@@ -1436,6 +1436,28 @@ def run_qwen2_5_omni(questions: list[str], modality: str):
     )
 
 
+# R-4B
+def run_r_vl(questions: list[str], modality: str) -> ModelRequestData:
+    assert modality == "image"
+    model_name = "YannQi/R-4B"
+
+    prompts = [
+        f"<|im_start|>user <image>\n{question}<|im_end|><|im_start|>assistant\n"
+        for question in questions
+    ]
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=16384,
+        limit_mm_per_prompt={modality: 1},
+    )
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
 # SkyworkR1V
 def run_skyworkr1v(questions: list[str], modality: str) -> ModelRequestData:
     assert modality == "image"
@@ -1622,6 +1644,7 @@ model_example_map = {
     "qwen2_vl": run_qwen2_vl,
     "qwen2_5_vl": run_qwen2_5_vl,
     "qwen2_5_omni": run_qwen2_5_omni,
+    "rvl": run_r_vl,
     "skywork_chat": run_skyworkr1v,
     "smolvlm": run_smolvlm,
     "step3": run_step3,