Yolov3 deleted the connected of darknet53.
What is connected like.
How many hidden layers are there? How many neurons are in the hidden layer.
all cnn connected the same?
Different hidden layers and output layers will be designed?
I will look forward to your patient answer.thanks
YOLO v3 is Darknet-53.
This is the first paragraph of Section 2.4 of the v3 paper…
We use a new network for performing feature extraction. Our new network is a hybrid approach between the network used in YOLOv2, Darknet-19, and that newfangled residual network stuff. Our network uses successive 3 × 3 and 1 × 1 convolutional layers but now has some shortcut connections as well and is significantly larger. It has 53 convolutional layers so we call it… wait for it… Darknet-53!
The architecture is depicted in Figure 1. of the paper…
Is there only one connected layer？1000 neurons?
For multi-class classification at a single scale, like what was done in v2 and discussed in this course, yes
The components preceding the AvgPool layer at the bottom of that Figure comprise the feature extractor. What you put downstream to perform detection depends on the task at hand. For multi-scale detection, the detection head and network output is substantially different and more complex.