-
作者您好,有个问题想请教您,就是我集成了bifpn,但是bifpn会对输入的底层特征图进行上采样以匹配上一层特征图的尺寸,但是当底层特征图的尺寸包含奇数时,会造成底层特征图上采样后尺寸与上一层特征图对不上的情况,比如大1的情况?您是如何解决的?是通过加载数据集时进行处理吗还是?期待您的回复
-
作者您好,请问您在EfficientDet中提出的BiFPN结构,在跨尺度连接的时候,是用的Concat操作(输出通道数为输入通道数之和)还是add操作(输出通道数与输入通道数一致)。希望有知道的朋友给我解答下,谢谢
-
请问,输入到BiFPN的p3_in~p7_in的channel数都必须一样,对吗?谢谢
-
I use backbone swin transfomer
-
你好,我在测试abcnetv2的时候报了这个错
Traceback (most recent call last):
File "demo/demo.py", line 75, in
demo = VisualizationDemo(cfg)
File "/root/code/adet/demo/predictor.py", line 40, in __init__
…
-
I saw the latest repo.Added the yolov4-p5,yolov4-p6.Can you explain some details about them?Such as mAP?and FPS?
What is PRN?
And About SE ,ASFF,biFPN,How to use these layer.How to modify the cfg?Ca…
-
大佬您好,请问有yolov5xP2CBAM-TPH-BiFPN-SPP的预训练权重吗?
-
Hi,
I am facing issue in biFpn module while starting the training using pretrained d1.
File "train.py", line 238, in
train()
File "train.py", line 183, in train
classification, …
-
In the class **BiFPN**, the layer **self.p4_down_channel** has the same definition with the layer **self.p4_down_channel_2**, same situation for **self.p5_down_channel** and **self.p5_down_channel_2**…
-
Thanks for sharing your code. When I loaded the code weights, I found that the dimensions were wrong, but I strictly followed your code to load.
`
def get. net():
nun_classes = 7
anc…