Hi, thanks for your great work!
I have a question regarding the implementation of EVA_VIT in the eva_vit.py file. It appears that the flash attention component from the StreamPETR code has been removed. Could you please explain if there was a specific reason for omitting it?
Hi, thanks for your great work!
I have a question regarding the implementation of EVA_VIT in the eva_vit.py file. It appears that the flash attention component from the StreamPETR code has been removed. Could you please explain if there was a specific reason for omitting it?