Stage 3: Act as a Service Provider

1.1 Objectives

Stage 3 will focus on testing deploying jobs from the Service Provider Dashboard.

1.2 Testing Type

This testing campaign will be focused on Integration testing. Verifying that different components or modules work together as expected for Service Providers.

1.3 Scope of Testing

The following features and functionalities will be tested during this campaign

• Installing the Service Provider Dashboard (SPD)

• Using the SPD and the wallet created on Stage 1 request to run a ML job (this will lock the mNTX funds)

• Waiting and checking to finish the ML job and review results

1.4 Testing schedule

This campaign will run from May 22nd and from next week will work with Stage 4(Acting as a Compute Provider). IMPORTANT Stage 4 Delay We are waiting until we can test / prove multiple jobs can be run by a service provider. Once we prove we have lots of jobs successfully running it will mean most users will have mNTX to claim. We expect at least one update to DMS and the SPD before we are at that point. We are expecting new releases of Compute provider dashboard next week. (Week commencing 22nd May)

2.1 Documentation

This section includes the relevant links to documentation needed for this test stage:

• If you haven't already, complete Stage 1 and Stage 2.

• Install Service Provider Dashboard.

• Enable debugging mode using the info in the troubleshooting tools section.

• Follow the steps to use the Service Provider Dashboard.

• Please post your successes or issues into the stage 3 Discord chat. Which will allow us to gauge how well things are working.

2.2 Monitoring and Reporting Tools

• Please log any problems you come across in the Issues Tracker

4.1 Get Involved

Please join our Discord Server and head to #role-assignment click on 4️⃣ to be given access to the relevant testing channels.

In order to receive mNTX your wallet you will need to be added to our Testnet Stage 4 Whitelist to receive mNTX, you can find a link to the whitelist in the Welcome group on Discord Stage 4 category.

4.2 NTX Incentives

• Testers will be rewarded for the following:

• Bug discovery	Critical bugs discovered

4.3 Communication Channels

• Stage 3 Category

• Stage 3 Forum

• Communication Channel for Stage 3 testers

If you have any issues or bugs to report please follow our procedure outlined in our Contribution Guidelines page on GitLab

Report will be added after testing is complete.

Video Walk Through

https://youtu.be/JkHLTqXeTZs

Step 1

You must have completed stages 1 and 2 so that you have a Pre-Production Testnet wallet with some mNTX, and have onboarded your device to NuNet (everyone has to allocate some resources to NuNet, even those ordering jobs as a service provider). Check that you have some DHT peers with the "nunet peer list" command before proceeding.

Step 2

Download and install the service provider dashboard and access it using localhost:9991 in your browser.

Instructions: https://docs.nunet.io/nunet-public-alpha-testnet/components-installation/service-provider-dashboard

Step 3

-Press ctrl+shift+i on localhost:9991 and click the Console tab so you can see that the process is working at all times.

-Connect your PreProd Tesnet wallet to localhost:9991

-Paste in this URL to the web app to order its machine learning script: https://gitlab.com/nunet/ml-on-gpu/ml-on-cpu-service/-/raw/develop/examples/cpu-ml-test-scikit-learn.py

-Choose CPU

-Choose Low

-Click "Please enter any additional required dependencies" and type in matplotlib

-Enter 30 (minutes) as the estimated time

-Click Next

Step 4

On the next page of the web app

-Check that the wallet address matches your own

-Enter an amount of mNTX to spend (1000 may work)

-Submit and sign the transaction in your wallet

Step 5

Wait for the job to complete. DO NOT refresh the web page, this will prevent you from seeing if the job completes. You should see in the Console (ctrl+shift+i) when the job completes.

Please report any errors or problems you have completing this process on the Gitlab issue submission page (with the exception of the issue stated below): https://gitlab.com/nunet/documentation/-/issues

IMPORTANT: It is currently a known issue that you cannot order a second job, even after the first one completes. Please only order one job and do not report that a second one cannot be ordered, this issue is already being fixed.