Classifying Images of Clouds in the Cloud with AutoML Vision

이 실습에서는 이미지를 Cloud 스토리지에 업로드하고 이 이미지를 사용하여
다양한 유형의 구름(적층, 적층 등)를 인식하도록 사용자 지정 모델을 교육합니다

Task 1. Set up AutoML Vision

Cloud AutoML API 활성화

메뉴

☰ > APIs & Services > Library

  • Cloud AutoML API 활성화 [Enable]
    image

Activate Cloud Shell

image

버킷을 만듭니다.

gsutil mb -p $DEVSHELL_PROJECT_ID \
-c regional \
-l us-central1 \
gs://$DEVSHELL_PROJECT_ID-vcm/

그리고 AutoML UI에 접속합니다.
https://console.cloud.google.com/vision/datasets

Task 2. Upload training images to Cloud Storage

메뉴

☰ > Cloud Storage > Browser
클라우드 트레이닝의 automl용의 이미지를 조금 전 만든 구글 스토리이제 복사합니다.

gsutil -m cp -r gs://cloud-training/automl-lab-clouds/* gs://$DEVSHELL_PROJECT_ID-vcm/

확인

gsutil ls gs://$DEVSHELL_PROJECT_ID-vcm/
gs://qwiklabs-gcp-01-84e906daed8e-vcm/data.csv
gs://qwiklabs-gcp-01-84e906daed8e-vcm/cirrus/
gs://qwiklabs-gcp-01-84e906daed8e-vcm/cumulonimbus/
gs://qwiklabs-gcp-01-84e906daed8e-vcm/cumulus/

Task 3. Create an AutoML Vision training dataset

gsutil cp gs://cloud-training/automl-lab-clouds/data.csv .
head --lines=10 data.csv
sed -i -e "s/placeholder/$DEVSHELL_PROJECT_ID-vcm/g" ./data.csv
head --lines=10 data.csv
gsutil cp ./data.csv gs://$DEVSHELL_PROJECT_ID-vcm/
gsutil ls gs://$DEVSHELL_PROJECT_ID-vcm/

image

  • New dataset > Single-label Classification > create dataset
  • Select a CSV file on Cloud Storage
  • gs:// [BrOWSE] 선택 합니다.

image

Task 4. Inspect the images

Dataset에서 Filter도 해보고, add label도 할 수 있습니다. image

image

Task 5. Train your model

[Train ] > [Start training] image

  • into the Set your budget: 8
  • Deploy model to 1 node after training: check

image

Task 6. Evaluate your model

[Evaluate]
Confidence threshold
image

Task 7. Generate predictions

[Test & Use]
Model > Upload images
image image image`