feat: add VLMEvalKit-compatible Qwen task variants for MMMU and MMStar (#1021)

Add new task variants that use VLMEvalKit-style prompt formatting:
- mmmu_val_qwen: Uses 'Question: {q}' prefix and 'Answer with the option letter only.' suffix
- mmstar_qwen: Uses same VLMEvalKit-compatible prompt structure

These variants help users reproduce benchmark scores closer to official Qwen
results reported in VLMEvalKit evaluations.

Usage:
  python -m lmms_eval --model qwen2_5_vl --tasks mmmu_val_qwen,mmstar_qwen ...

Addresses score reproduction gaps reported in Issues #935, #932, #881, #901

Author: Luodian

lmms_eval/tasks/mmmu/mmmu_val_qwen.yaml

@@ -0,0 +1,22 @@
+dataset_path: lmms-lab/MMMU
+task: "mmmu_val_qwen"
+test_split: validation
+output_type: generate_until
+doc_to_visual: !function utils.mmmu_doc_to_visual
+doc_to_text: !function utils.mmmu_doc_to_text
+doc_to_target: "answer"
+doc_to_messages: !function utils.mmmu_doc_to_messages
+process_results: !function utils.mmmu_process_results
+
+metric_list:
+  - metric: mmmu_acc
+    aggregation: !function utils.mmmu_aggregate_results
+    higher_is_better: true
+
+lmms_eval_specific_kwargs:
+  default:
+    format: "qwen3_vl"
+    pre_prompt: "Question: "
+    post_prompt: "Answer with the option letter only."
+    open_ended_prompt: "Please answer the question directly."
+include: _default_template_yaml

lmms_eval/tasks/mmstar/mmstar_qwen.yaml

@@ -0,0 +1,11 @@
+dataset_path: Lin-Chen/MMStar
+task: "mmstar_qwen"
+doc_to_visual: !function utils.mmstar_doc_to_visual
+doc_to_text: !function utils.mmstar_doc_to_text
+process_results: !function utils.mmstar_process_results
+
+lmms_eval_specific_kwargs:
+  default:
+    pre_prompt: "Question: "
+    post_prompt: "Answer with the option letter only."
+include: _default_template_yaml