Problem statement : We have to create a machine learning project and automate this with the help of Devops .
A docker instance will run my ml code which will be downloaded from the github. We will train a machine learning model in the docker instance , after training we test the accuracy of the model and notify the developer through email. If the accuracy of the model is less than 80% , we have to tweak our code and retrain the model. All the task should be monitored.
To achieve this setup we will use jenkins as our automation tool.
We have to follow the steps to achive this task :
Step 1: create a dockerfile which will launch the docker instance
Step 2:Create a job to download github repository
Step 3: Identify the type of ml code i.e. it is deeplearning code or regression code etc and the launch the container with respective software . In my case it is cnn model.
Step 4: Check for accuracy
Step 5:Tweak the ml code if the desired accuracy is not achieved
Step 6:Notify the developer about the accuracy of the code
import smtplibs = smtplib.SMTP(‘smtp.gmail.com’, 587)
#Message that will be sent to
mailmsg = “accuracy is achieved” s.sendmail(“firstname.lastname@example.org”, “email@example.com”, msg)
Here is my link to github: https://github.com/gyaneshsharma/mlopsproject1
please give your valuable suggestion about the project.
This is my first ever blog so please give your valuable feedback.
I would like to thank my mentor Mr Vimal Daga for guiding me throughout my learning journey of latest Technology.