How to manage ML datasets with Vertex AI

[@salonitilekar4889]

Very well explained!

[@andrewsuttar]

Priyanka! BEST instructions EVER from Google - normally I find google help resources TERRIBLY scaffolded but this was AWESOME - well done!

[@obscurecloud5575]

Very interested in how Vertex AI can help with graph ml/embedding tasks

[@geminiexpertAi]

Thanks this was a great lesson with detailed description at work How can I get more detailed lessons on this topic?

[@sugunakarvure]

Thank you for your Informative lecture. I have a dataset related to healthcare management, and I am interested in building one using V-AI. Are there any customised models? I would like to listen

[@amitpawar1677]

Good and informative video Priyanka

[@azadansary-f8c]

Can we create a dataset form .txt(or pdf/doc) files and train the model?

[@shikhardev1]

When you create datasets, for example if create an image dataset sourcing images from cloud storage, does it create a new physical copy?

[@nexai_official]

well introduction! thanks for sharing!

[@archanapereira1333]

Thanks Priyanka. How abt the data versioning in Vertex AI ?

[@moraholguin]

Hi, do you have an example on how to connect a REST API to a google Cloud Storage Bucket?

[@TreySpeakman-b2n]

Is there any documentation on this earth about formatting the csv import file. The two seconds from this clip is all I have been able to find.

[@manjureddy5606]

Is there a codelab which explains accessing the AUTOML endpoint to query the model using cloud functions ?

[@furry2fun]

bro if u find it please reply

[@lengmobile]

Would like to see tools for conversion/edit to jsonl. Dash board region selection allowed Singapore region, but datasets upload only allow US/EU selection.

[@vichakshansuresh]

Thanks Priyanka. Just enabled the API. Is there "DataPrep" inside Vertex AI?. How do I "code" (instead of clicking UIs) in vertex?

[@rexmane11]

Notebooks is part of vertex

[@greendsnow]

Do you think they draw the dots of the i as ♥ for a subliminal message?

[@rajmahalingam3936]

I am not able to see my Bigquery Data sets in tabular model create dataset, I did provide the necessary permission, can you tel lwhich permission to give and to what service account, how much time does it take to take effect?

[@newcooldiscoveries5711]

Thank you! It was very helpful. :)

[@allandclive]

Can you create a tutorial on creating Labelled CSV?

[@rishabhgupta-he2wg]

My datasets involves using audio. Can I use vertex AI? If yes then how?

[@jincyquones]

Spectrograms

[@patmclaughlin107]

Isn't forecasting just another type of regression problem?

[@parisfrancepp]

I had the same question. I found out that forecasting is a kind of regression but more used for predicting future values, whereas regression can predict past, current or future values.

[@patmclaughlin107]

@@parisfrancepp But both are about predicting continuous numbers, right? How does the model know or care whether the values are from the past, present or future?

[@parisfrancepp]

@@patmclaughlin107 Yes they both predict continuous numbers but keep in mind this key difference: Regression predicts a single value, while forecasting predicts a sequence of values (a time series).
The target value of a regression is a SINGLE value that depends only on the feature columns values. The target valueS of a forecasting depends on the feature columns values AND on the context

[@patmclaughlin107]

@@parisfrancepp Understood. Can you please clarify what "context" means? Thank you.

[@parisfrancepp]

@@patmclaughlin107 by context I mean the historical data and patterns of your time serie.
Ex with a stock price:
At a time t, the stock price depends on the feature columns values like the revenue of the company, the profitability, the P&L.. but it also depends on the historical trends of the stock (if it has been falling for the last 3 years, it is likely it will keep falling, this is the historical pattern =context)

[@sanjoynathgeometrifyingtrigono]

How much will i have to pay to google?

[@尾崎元恒-j2u]

5/25の0時が作戦開始時間だと考えてます。

[@DibizURL]

well actually how to create my own datasets?

[@navinmp2373]

👍

[@陈泽-t5m]

你们产品的注册选项里面我看到了你们有选择的尊重，尊重其他国家是自我修养的体现，如果你们觉得自己被尊重不是你们的修养有多好，而是我的

[@greatpose5423]

Bakbas

[@alawadhi5257]

useless