Hey Readers, Today, I am here with another blog. We will learn something special about creating an AL/ML model using Images. To perform this operation, Knoldus made a project which is known as Studio9.
First, we will know what is studio 9?
Studio 9 platform Enables developers to create, train and test the ML models quickly with a single click of a button along with other features like pipeline and custom pipeline creation using Jupiter lab.
In studio 9, We can create a model using images, CSV, and video type data. So in this blog, We will see how we can create a model using video. To perform this operation, We need a few video data. If we are creating a new model then we need to have two videos – train and test because We will train a model on the basis of video data which is known as trained data. After the creation of a new model, We can train any other data on the basis of that model. I have divided this process into steps.
You can follow these steps to create a model:-
You need to log in to studio9. If you have a login ID otherwise you can click on the sign-in button and fill in the required details and create a user for studio9.
Step 2: Create an Album and upload the data – Train and test data:
Just simply click create Album on the left-side and fill in the required details as per the requirement as below:
Actually, You can upload the data from the AWS s3 and also provide the access, secret key, and session token. You can use the below command to generate token and keys:
aws sts assume-role --role-arn <Arn of your s3> --role-session-name s3-access --profile <Your aws configure profile name> --duration-seconds 20000
aws sts assume-role --role-arn arn:aws:iam::25235823423:role/s3-access --role-session-name s3-access --profile project --duration-seconds 20000
Now click on the upload button. We need to follow the same step to upload test data.
Step 3: Now You can create an experiment using these data.
But you think that we have a small data then we can apply the Augmentation to the data. It will generate more pic from the existing data.
Navigate to the visualization option on top and click on Apply data augmentation. You will see two options to fill. One for the output Album name and the second for selecting the input album. You can perform this operation on both data – train and test.
Now, You are ready to create an experiment. Navigate to the lab section on
the top page.
Now, You can fill in the required details like experiment name, experiment
type – We will select CV TL train because We are using image data so We
will select Computer vision, Training mode – We can select 1 step or 2
steps as per requirement, UTLP- select an algorithm, Model type
- We will select a classifier because we are using image data. And upload
the train and test data.
Now you will see an Advanced option just below the train data option. If you
check the Enable Automated Data Augmentation
Then It will generate more images by applying augmentation.
In the last just click on Create.
Hurry!!, Your CV model is created. Now you can use this model to train
other data. You can find your model in the cv model option on the left side.
Studio9 is an open source platform for doing collaborative Data Management & AI/ML anywhere Whether your data is trapped in silos or you’re generating data at the edge, Studio9 gives you the flexibility to create AI and data engineering pipelines wherever your data is. And you can share your AI, Data, and Pipelines with anyone anywhere. With Studio9, you can achieve newfound agility to effortlessly move between computing environments, while all your data and your work replicate automatically to wherever you want.
If you want to create a model from the video. We made this quite simple. You just need to upload two videos – test and train. our system will generate the image from the video and then you can perform the same operation as we performed on the image.