Google Fly Cup Challenge: Pilot

Hi friends I am Keshav from TECH_ED let's solve this Pilot one too! 

Task 1

Starting off with this command :

gsutil mb -l [REGION] gs://[PROJECT-ID]

 

Task 2

 

BigQuery > Create "drl" dataset > Create Table "pilot-event-times"

 

Cloud Storage : spls/gsp396/lab2.csv

 

Schema : auto-detect

(Watch Video for demonstration, TECH_ED)

 

Task 3

 

Vertex AI > Workbench > JUPYTERLAB > VAI notebook > Terminal 

 

gsutil cp gs://spls/gsp396/DRL2.ipynb  .

(Watch Video for demonstration, TECH_ED)

 

Task 4

 

Run all -> DRL2.ipynb

 

PROJECT variable = project id

 

(Watch Video for demonstration, TECH_ED)

Task 5

 

Vertex AI > Model Registry > Import > model "drl-pilot-ranking"

 

<choose As per the lab>

 

Model Settings > Model Version 2.8 

 

Location: gs://[project-id]/model

 

(Watch Video for demonstration, TECH_ED)

 

 

Task 6

 

Vertex AI > Create Endpoints "drl-pr-endpoint"

 

<region as per the lab>

 

Advanced scaling settings = n1-standard-2 

 

(Watch Video for demonstration, TECH_ED)

 

 

Task 7

 

Add New cell in Notebook :

 

 

from google.cloud import aiplatform

 

PROJECT = '[project-id]'

LOCATION = '[region]'

ENDPOINT_ID = '[your-endpoint-id]'

 

endpoint = aiplatform.Endpoint(ENDPOINT_ID, project=PROJECT, location=LOCATION)

 

instances = [{'pilot_id': [11],

          'ename': ['Practice'],

          'events_name':['Project Manhattan'],

          'minimum_time': [80.0]

         }]

 

endpoint.predict(instances = instances)

 

Now Follow as doing in the video.

Congrats, the last task is also done, you can comment your doubts in the comment section of video. If you find it little bit helpful please support by subscribing the channel TECH_ED for more such content.