Comments (7)
The time to execute the actual job should not take more than 2-5 minutes, however it can potentially take an equivalent time to acquire the VMs for the cluster. Note the comment just under where you specify the node and worker types.
If you have the permissions to do so I would set up an instance pool and supply the pool ID to the commented out driver_instance_pool_id
and instance_pool_id
attributes. This means that for each job you can acquire resources from that warm pool of instances. Note that you will then have to comment out node_type_id
and driver_node_type_id
.
Your deployment.yml
would look something then like the following:
from e2e-mlops.
Creating a starter job is not required. The first job you submit against the pool may take a few mins to acquire those VMs for the pool. But then that pool will remain warm to the time period you have set. Subsequent jobs which use this pool will then take a shorter time to acquire resources as they will acquire VMs from the now warm pool.
from e2e-mlops.
It looks like you're running on an Azure workspace, but supplying an AWS instance type.
In deployment.yml
we define node_type_id
and driver_node_type_id
, using i3.xlarge
instance types. These are AWS-specific. Ultimately, the choice of instance type is up to the user, but the nearest equivalent on Azure would be Standard_DS3_v2
. I would suggest replacing i3.xlarge
with Standard_DS3_v2
and give it a go.
from e2e-mlops.
Thanks @niall-turbitt . Setting the node_type_id
and driver_node_type_id
to Standard_DS3_v2
solves it. Now the launch
command can be executed. The speed for running this job seems quite slow though. Would using a larger instance likely to help? Like Standard_DS4_v2
or Standard_DS5_v2
?
from e2e-mlops.
Thanks. Just to double check, what should be the compatible preloaded Databricks Runtime version? Should it be 11.0 ML (Scala 2.12, Spark 3.3.0)
or 11.0 (Scala 2.12, Spark 3.3.0)
?
from e2e-mlops.
You should select DBR 11.0 ML (Scala 2.12, Spark 3.3.0)
from e2e-mlops.
Thanks! I have created the pool now. Do I also need to add a starter job as the video explained? However, the interface is quite different. I couldn't find a place to specify the use of pool here.
Is there any sample code I could copy over for the starter job?
Thank you.
from e2e-mlops.
Related Issues (6)
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from e2e-mlops.