Hi !
I didn’t found this online, so I decided to write it down. How to create a dataset with images to be used on the Azure Machine Learning designer, when training a model using the DenseNet prefab and samples.
Let’s start defining which Azure Storage Account we will use. I’ll use the default one. So, I navigate to the DataStores section, and open the default datastore.
In the Account Name section, we can navigate the Storage Account.
Now let’s browse the Storage Account and create a new folder in the Blob container. I’ll name the folder “squirreltraining01”
Inside the container, we need to create diferent directories for our images. Each directory will be later used as the label for the training data set.
In the following example, I added a directory named “squirrel” and added 50 images to be used as squirrels.
After uploading the Space Wolves training images in a new directory, the Storage account is ready to be used.
Now we need to go back to the Datastores section and create a new datastore. Once we select the previous storage account, we can also select the blob container that we used to upload all of our images.
The new datastore is ready to be used !
We are getting closed, now let’s navigate the Datasets section and create a new dataset from datastore.
In the [Basic Info] section, we define the name and select the [Dataset Type] as [File].
In the [Datastore selection] we choose the datastore that we create previously and define the general path [/**] to include everything.
And that’s it. Our new dataset is created and we can use it from Azure Machine Learning Designer projects.
Tomorrow I’ll show how to use this new dataset in a Azure ML Designer experiment.
Happy coding!
Greetings
El Bruno
More posts in my blog ElBruno.com.
More info in https://beacons.ai/elbruno
¿Con ganas de ponerte al día?
En Lemoncode te ofrecemos formación online impartida por profesionales que se baten el cobre en consultoría:
- Si tienes ganas de ponerte al día con Front End (ES6, Typescript, React, Angular, Vuejs…) te recomendamos nuestros Máster Front End: https://lemoncode.net/master-frontend#inicio-banner
- Si te quieres poner al día en Backend (stacks .net y nodejs), te aconsejamos nuestro Bootcamp Backend: https://lemoncode.net/bootcamp-backend#bootcamp-backend/banner
- Y si tienes ganas de meterte con Docker, Kubernetes, CI/CD…, tenemos nuestro Bootcamp Devops: https://lemoncode.net/bootcamp-devops#bootcamp-devops/inicio
El Bruno 