Support for Qwen3VL Evaluation

Could you please add support for Qwen3VL model evaluation? Qwen3VL is a widely used multimodal model, and we need to run standard evaluation benchmarks (such as MME, SEED-Bench, POPE, etc.) on it. Currently, this framework does not support Qwen3VL yet. It would be greatly appreciated if you could: Add official support for Qwen3VL Provide a config example for Qwen3VL evaluation Let us know the estimated timeline if you have a plan Thank you very much!