Hello, it's a very cool model. It would be very nice if some multimodal model ( e.g. Qwen2-VL) could be trained as well.
· Sign up or log in to comment
