Regarding the issue of improving the model

[josieisjose]

@GuangxingHan Hello, I have recently been training my dataset using your FCT model. I found that my dataset achieved very good results on your model with shots=3, 5, 10, all with an AP above 80. Consequently, I attempted to make some innovative changes to your model. After numerous trials, I found that the AP dropped to the 20s. I am quite puzzled by this and would like to know if you have encountered a similar situation during your experiments?