In this video we have explained how to train YOLO v4 for custom object detection on google colab utilizing the free GPU resources. This video is very special because it provides complete overview of changing the make file configuration file and crating training and testing dataset feel free to add your custom class and train your own model. we have also explained how to use trained model to detect object on live video.
This video cover:
Subscribe : https://www.youtube.com/channel/UCoB1kaaQrQEkB-ud2QxmgUQ
Add crome extension to download images by below URL
https://bit.ly/3nXfPI8
Download labelImg tool with below link
https://tzutalin.github.io/labelImg/
git link to clone darknet on colab
https://github.com/pjreddie/darknet
Get train and test data generator from here
https://github.com/jakkcoder/training_yolo_custom_object_detection_files
Note for point 4 :- I am not the authors for 2 py files complete credit goes to authors for
creating-files-data-and-name, creating-train-and-test-txt-files files.
Download pre-trained weights for the convolutional layers (154 MB):
http://pjreddie.com/media/files/darkn…
command to train the model (take care of single line and spaces)
!darknet/darknet detector train custom_data/labelled_data.data darknet/cfg/yolov3_custom.cfg
custom_weight/darknet53.conv.74 -dont_show
#yolo #tensorflow #deep-learning