Question about the availability of fine-tuned policy checkpoints for HEAR-Bench

[Zhengxian-Huang]

Hi authors,

Thanks for your great work!

I noticed that the repository provides links to download several HEAR-related model components from Hugging Face, such as HEAR-Qwen3-Omni-30B-A3B-Instruct-Pruned, HEAR-Qwen3-0.6B, and HEAR-mimi.

However, from the current documentation, it seems that running evaluation on HEAR-Bench may require a fine-tuned policy checkpoint. I wonder whether the fine-tuned policy checkpoints used for HEAR-Bench evaluation are planned to be released as well. And if yes, any time schedule?

Looking forward to your reply!