after I run the traindetector.py, I can get the vg{epoch}.tar
base on pretrained detector { 'vg_{epoch}.tar' } , Do I have to run 'train_models_sgcls.sh 2' scripts?
I am confused to use the checkpoint file of a type of 'tar'.
Is that right to just use the 'tar' checkpoint itself?
after I run the traindetector.py, I can get the vg{epoch}.tar base on pretrained detector { 'vg_{epoch}.tar' } , Do I have to run 'train_models_sgcls.sh 2' scripts? I am confused to use the checkpoint file of a type of 'tar'. Is that right to just use the 'tar' checkpoint itself?