Qwiklabs — Classify Images of Clouds in the Cloud with AutoML Vision
Qwiklab : https://google.qwiklabs.com/focuses/8406?parent=catalog
Enable Cloud AutoML API
Create Google Bucket
export PROJECT_ID=$DEVSHELL_PROJECT_ID
export QWIKLABS_USERNAME=<USERNAME>gsutil mb -p $PROJECT_ID -c standard -l us-centrall gs://$PROJECT_ID-vcm
you can find this on UI
copy image data from google provide storage location
export BUCKET=$PROJECT_ID-vcm
gsutil -m cp -r gs://automl-codelab-clouds/* gs://${BUCKET}
You can find there are free folder show up on bucket, and each of them have image data.
Download the meta data information, we will identify which picture will belong to which label
gsutil cp gs://automl-codelab-metadata/data.csv .
head data.csv
Change bucket name
sed -i -e "s/placeholder/${BUCKET}/g" ./data.csv
Copy metadata file to our bucket
gsutil cp ./data.csv gs://${BUCKET}
Now we have the data, we can start to build AutoML model
On the search bar type AutoML and select Vision service
First, We need to create dataset
It will take a while to loading images.
After completely, you can check the image data.
Click the Label STATUS, you can look at the basic information
Now click the train button, we can start to train our model.
Starting training…..
After training complete, you can go to EVALUATE part, look at the result matrices
And you can click TEST button, use some data to do testing