Project 6d - Object detection II, training with Google Colab

Project start date: 27 February 2022

Project finish: 28 February 2022

Objective: To calculate the weights file for the license plate detection program using Darknet in Colab with GPU.

Create a folder called project_6d in the myColab folder.

Copy the attached Jupyter Notebook into the /myColab/project_6d folder.

project_6d_V0.ipynb.zip

Download ZIP • 87KB

Open the notebook and click "Runtime", then "Change runtime type", "GPU" under "Hardware accelerator" text and then "save" in the bottom right had corner.

Remember that when you run the Jupyter-Notebook, Google will ask for some permissions that you must accept.

Results are as follows:

a) weight files which are written to /cfg/backup:

myYolov4_1000.weights

myYolov4_2000.weights

myYolov4_3000.weights

myYolov4_4000.weights

myYolov4_last.weights

myYolov4_final.weights

myYolov4_best.weights

b) MAP (Mean Average Precision) chart which is a png file written to /cfg/chart.png

If you have any problems or need clarification please contact me: jscanova@gmail.com