Test dates for azure ml

Question

I created data sets of various sizes: 1GB, 2GB, 3GB, 4GB (<10 GB) and running various machine learning models on Azure ML.

1) Can I find out what the characteristics of the server (RAM, CPU) are that are provided in the Azure ML service.

2) Also, sometimes the reader says “Memory exhaust” for> 4 GB of data. Although azure ml should be able to process 10 GB of data in accordance with the documentation.

3) If I perform several experiments (in different browser tabs) in parallel, it takes longer.

4) Is there a way to install RAM, CPU cores in Azure ML

Answer 1

I have a partial answer: 1. no, he is distracted

2. The following data types can expand into larger datasets during normalization and are limited to less than 10 GB:

   Sparse categorical strings Binary data

(see this )

3. I'm not sure, but while working on this, I did not experience any changes when starting one experiment and several experiments

4. you can scale cars in a standard tier (see this )

Answer 2

I would recommend taking a look at the new "Visual Interface" for the Azure ML service, which allows you to overcome the 10 gigabyte limit and create your own computing clusters.

// Video about the BUILD 2019 announcement: https://www.youtube.com/watch?v=QBPCaZo9xx0