HEUzhouhanwen
HEUzhouhanwen
The loss is stable at 30000-40000,I think it is related to Pretraining!
Hello Try evaluating the validation set.I use ../robot-grasp-detection/models/grasp/m4/m4.ckpt, but the accuracy is about 30%, what is my problem? Thank you!
I see! Thank you!
But I still don not understand why the loss is stable at 30000-40000!
I met the same error like you ,I cannot get the bounding box on RPI 3 too,I‘m sure there is no problem with the program,because it works well on pc...
@Nyceane I solve the problem ,I try all the way to find the error but failed.Today I notice the pc(it can get the bounding box) use the previous version of...