[Test]update accuracy test of models (#4911)

### What this PR does / why we need it?
Delete accuracy tests for models that are no longer retained：
- Meta-Llama-3.1-8B-Instruct
- llava-1.5-7b-hf
- InternVL2-8B.yaml
- InternVL2_5-8B.yaml
- InternVL3-8B.yaml

Add accuracy tests for the new models：
- Llama-3.2-3B-Instruct
- llava-onevision-qwen2-0.5b-ov-hf
- Qwen3-VL-30B-A3B-Instruct

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: MrZ20 <2609716663@qq.com>

.github/workflows/nightly_test_a2.yaml

@@ -86,15 +86,13 @@ jobs:
               - Qwen3-8B-W8A8
               - Qwen3-VL-8B-Instruct
               - Qwen2.5-Omni-7B
-              - Meta-Llama-3.1-8B-Instruct
           - os: linux-aarch64-a2-1
             model_list:
               - ERNIE-4.5-21B-A3B-PT
-              - gemma-3-4b-it
-              - internlm-7b
               - InternVL3_5-8B-hf
-              - llava-1.5-7b-hf
               - Molmo-7B-D-0924
+              - Llama-3.2-3B-Instruct
+              - llava-onevision-qwen2-0.5b-ov-hf
           - os: linux-aarch64-a2-2
             model_list:
               - Qwen3-30B-A3B
@@ -103,6 +101,7 @@ jobs:
           - os: linux-aarch64-a2-4
             model_list:
               - Qwen3-Next-80B-A3B-Instruct
+              - Qwen3-VL-30B-A3B-Instruct
     uses: ./.github/workflows/_e2e_nightly_single_node_models.yaml
     with:
       vllm: v0.12.0