You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, thanks for your great work. I used my own datasets to train the model, but the acc is very low. When I use resnet-50/convnext-tiny as backbone to train, the acc can achieve at least 90%+, but use the inceptionnext as backbone, the acc is only 67.9% as following images
And this is my train scripts as following
since of my class number is only 4, I change and add some code as followings
Could you please tell me why this happen and give me some advice ?
The text was updated successfully, but these errors were encountered:
Hi, thanks for your great work. I used my own datasets to train the model, but the acc is very low. When I use resnet-50/convnext-tiny as backbone to train, the acc can achieve at least 90%+, but use the inceptionnext as backbone, the acc is only 67.9% as following images
And this is my train scripts as following
since of my class number is only 4, I change and add some code as followings
Could you please tell me why this happen and give me some advice ?
The text was updated successfully, but these errors were encountered: