Evaluation based on the qwen model

[zhoujixi02-prog]

I noticed on Huggingface that Navila also has checkpoints based on Qwen. Would you release the Qwen-based benchmark code? Are there any Qwen-based benchmark results? Thank you very much.