Hi Sir...Really Insightful set of videos/sessions...Thanks a lot for your contribution to the Community...Request to create a video on identifying duplicate records from multiple data sources/tables..using Supervised/Unsupervised/recordlinkage comparisons..It would be really helpful 🙏.. Thanks, Vijay
Thanks for the demo. Here we just have 3 features, in our scenario over 10k features are created so this there a way to provide the list to feature view?
I didn't really understand. how it is useful. Data can be different in real time both training and prediction pipeline data then how can we use feast there?
Is created time stamp and event time stamp must in all datasets which we are gonna work on with feast? Is the dataset is not having those, how to add those.. Or is there any other way?
@@AIEngineeringLife thanks for the response.. Like firestore what other online stores are supported by feast as of now? 2) how about minio s3 for registry?
If you want to learn MLOps here is my playlist - ua-cam.com/play/PL3N9eeOlCrP5a6OA473MA4KnOXWnUyV_J.html
Kindly order the playlist so that it is helpful to us
Hi Sir...Really Insightful set of videos/sessions...Thanks a lot for your contribution to the Community...Request to create a video on identifying duplicate records from multiple data sources/tables..using Supervised/Unsupervised/recordlinkage comparisons..It would be really helpful 🙏..
Thanks,
Vijay
Code used in this video is here - github.com/srivatsan88/UA-camLI/tree/master/feast
Thanks for the demo. Here we just have 3 features, in our scenario over 10k features are created so this there a way to provide the list to feature view?
I didn't really understand. how it is useful. Data can be different in real time both training and prediction pipeline data then how can we use feast there?
is it necesary for bigquery table to have timestamp columns in it. The one we reffer for data source?
Is created time stamp and event time stamp must in all datasets which we are gonna work on with feast? Is the dataset is not having those, how to add those.. Or is there any other way?
how can feast read/extract features from a kafka stream?
Thanks for the demo. Can we use postgres or any rdbms instead of gcp dbs and also for registry can we use object store like minio S3?
Shabarish.. Yes it can be anything of your choice. Currently Feast does not support rdbms though natively to my knowledge
@@AIEngineeringLife thanks for the response.. Like firestore what other online stores are supported by feast as of now?
2) how about minio s3 for registry?
great presentation can u pls share the note book sir
It is in the comment I have added below the video
@@AIEngineeringLife thk u sir
@@AIEngineeringLife wen do we go for kubw flow with feast sir pls suggest
Timely and great tutorial! If I may ask, do you use this library (in your job) for serving features to production models?
Henok.. We have built our own system few years back as at that time many such products was not available. We recently evaluated this or a use case
How to automatically identify features used by ML Model?
Please do a video on model deployment using seldon core
Will try to do it Chandra in few months