hello, I have seen in your tech report that expr&pose is trained using full-paramter tuning, I wonder if it can be trained using lora? Have you tried?
hello, I have seen in your tech report that expr&pose is trained using full-paramter tuning, I wonder if it can be trained using lora? Have you tried?