Opinie ([DEPRECATED] Structured data prediction using Vertex AI Platform)

Opening up the jupyterlab was very slow. I wasn't sure if the terminal was working or not because of how long it actually took to start working

Christopher L. · Sprawdzono około 4 lata temu

Need to activate "Notebook API" for vertex AI

Daniel D. · Sprawdzono około 4 lata temu

Sean R. · Sprawdzono około 4 lata temu

Luke C. · Sprawdzono około 4 lata temu

Alchemical R. · Sprawdzono około 4 lata temu

When checking the directory structure both inputs produced "CommandException: One or more URLs matched no objects." And I had run into an issue with region, my notebook was in us-west1-b, when i deployed a trained model the region was not available.

Tony R. · Sprawdzono około 4 lata temu

Jesus M. · Sprawdzono około 4 lata temu

Pierre-Alexandre D. · Sprawdzono około 4 lata temu

P S. · Sprawdzono około 4 lata temu

Reshu R. · Sprawdzono około 4 lata temu

Karan R. · Sprawdzono około 4 lata temu

BO QUAN PATRICK L. · Sprawdzono około 4 lata temu

missing instruction, needed to tweak things to get the notebook run and train the model, the last part on deploying the model did not work so this lab needed more fixed on it.

Elis T. · Sprawdzono około 4 lata temu

Pretty clear, quite the same as End-to-End ML, but no real practice of Vertex

Axel D. · Sprawdzono około 4 lata temu

Jeyaram G. · Sprawdzono około 4 lata temu

Andreas B. · Sprawdzono około 4 lata temu

The command line are still using the old AI platform, not the Vertex AI.

Lingchuan C. · Sprawdzono około 4 lata temu

Eduardo M. · Sprawdzono około 4 lata temu

The lab was titled "Structured data prediction using Vertex AI", but the only VertexAI-oriented detail was creating a notebook in Vertex AI — which, btw, is virtually the same as in Google AI Platform. This clickbait-y nature is even reflected in the summary: "how to trained, evaluate, and deploy a machine learning model in Vertex AI Notebooks" (note how they replaced the word "Platform" with just "Notebooks"). The lab is so not-Vertex-AI-oriented, even the created jobs and models are not displayed in Vertex AI's menus, but in AI Platform's, where one is greeted with the following message: "This is AI Platform. Migrate your resources to Vertex AI to get the latest features like data labeling, AutoML-powered training, and endpoint management." How ironic ;). --- BTW, what is the sense of checking the existence of the same files in two consecutive commands? Is there a possibility that the created CSV files disappear instantaneously after a single `gsutil ls`? The commands, for reference: """ Verify CSV creation Verify that we correctly created the CSV files in our bucket. ``` %%bash gsutil ls gs://${BUCKET}/babyweight/data/*.csv ``` Check data exists Verify that you previously created CSV files we'll be using for training and evaluation. ``` %%bash gsutil ls gs://${BUCKET}/babyweight/data/*000000000000.csv ``` """

Mateusz Z. · Sprawdzono około 4 lata temu

Tarun B. · Sprawdzono około 4 lata temu

I wasn't able to complete the task since some error related to access to Google Storage Bucket showed up

Mateusz K. · Sprawdzono około 4 lata temu

Muhannad B. · Sprawdzono około 4 lata temu

José Luis G. · Sprawdzono około 4 lata temu

5 stars = Very satisfied

Mrunal H. · Sprawdzono około 4 lata temu

Paul B. · Sprawdzono około 4 lata temu