Pretrained projector weights for AudioVisualText task

[Freecheck100]

Hi there,

May I ask if it is possible to release the pretrained projector weights for the various LLMs(Qwen2, Qwen2.5-VL, LLaMA3) on the AudioVisualText task? The computational resources required for pre-training are prohibitively high for us to reproduce.

Thanks!

[CROVO1026]

I need these too