@xingyizhou Hi, i am sorry to bother you, my question is:I have changed the backbone of your model,but i don't know how to make sure i have get the highest mAP of my model. In other issues you had explained that model_best is not the model which has highest mAP and instead it is model_last,so in your mean visualization train loss is not useful, so what should i do to make sure i have get the highest mAP,train more epoch?
Train more epochs should be the most straightforward way. Unfortunately I can not give further suggestions without knowing more details about your model.
@xingyizhou Hi, i am sorry to bother you, my question is:I have changed the backbone of your model,but i don't know how to make sure i have get the highest mAP of my model. In other issues you had explained that model_best is not the model which has highest mAP and instead it is model_last,so in your mean visualization train loss is not useful, so what should i do to make sure i have get the highest mAP,train more epoch?
thanks for your answer in advance!