0

I am pretty new to ML and my question may look strange. Especially the last part of it.

1)As far as I understand Darknet53 is an integral part of Yolo just as Resnet50 is a part of R-CNN Am I right?

2)On the other hand I understand that the R-CNN alternatively can be used based on VGG instead of the Resnet, 16 am I right?

3)Can YOLO also be used based on something else but Darknet53 ?

4)If not what was the use in defining Darknet53 as not integral part of YOLO but as some separate section with its own name? Maybe it has some non-Yolo use?

5)I have almost finished an online course in YOLO-V3\4 and I am quite sure that in this course they do not use nor Pytorch nor Tensorflow nor Keras, but when I read online tutorials they often do mention at least one of them relating to YOLO. So what do they actually do? And what is their advantage comparing to not using them at all as in a course I am doing now? Thank you very much.

Igor
  • 181
  • 10

1 Answers1

2

Lets start by listing what is what.

  • RCNN : Is a type of CNN Model
  • Resnet50, DarkNet53 and VGG : Are implementations of a CNN Model

Now moving to your questions.

  1. Yes Darknet53 is the backbone of Yolo3. No, Resnet50 is not part of RCNN. Its just one implementation of a CNN model.

  2. Yes

  3. Yes

  4. N/A

  5. Pytorch,Tensorflow and Keras are frameworks for ML. They make it easier to implement and play around with neural networks. You of course do not "Need" them and can write everything from scratch. Using frameworks just makes your life easier.

Marib Sultan
  • 76
  • 3
  • Thank You sir. I start getting (hopefully) the terminology. About question 5: In the the forum of the online course I am taking, teacher answered as follows about YOLOv3\4 - to RUN pretrained model I can use : CV2 or Darknet OR TENSORFLOW... - to TRAIN it I can use: Darknet OR TENSORFLOW Does it make sense? I am confused since I thought That DARKNET is also a type of CNN, how can it be replaces with TENSORFLOW which is framework? I miss something very important' please help me fill the blanks in the taxonomy of terms and subterms . Thank you very much. – Igor Aug 09 '22 at 05:56
  • 1
    I see you are probably confused between Darknet and Darknet53. Darknet is a framework for neural networks just like Tensorflow. While Darknet53 is one implementation of a CNN model used as Yolo's Backbone. Darknet (the framework) is popular because the original Yolo implementations were in Darknet framework. – Marib Sultan Aug 09 '22 at 15:39
  • Now I See. Thank you very much! – Igor Aug 09 '22 at 19:32