Explain it is necessary to have some peers to request to run a ML job
Description
Click to expand
Who
- @avimanyu786 -- developer
- @janaina.senna -- reviewer
What
- Add on service-provider-dashboard readme file, to request to run a ML job only when there is a list of at least two peers on the DHT peers section of
nunet peer listcommand. - Change the example model to this one
https://gitlab.com/nunet/ml-on-gpu/ml-on-cpu-service/-/raw/develop/examples/cpu-ml-test-scikit-learn.pythat will run faster in a machine with low resources.
How
- Change the service-provider-dashboard readme file on gitlab and also on gitbook since we are still working on putting the readme files automatically on gitbook.
Why
- Analyzing the issues opened by community members here, we need to add on documentation to request to run a ML job only when we have peers on DHT section; the more peers the better (it seems six is a good number). If not, the user will probably get a
404 Not Founderror: This error occurs when the DHT (Distributed Hash Table) does not have any peers with matching specs. After analyzing and filtering machines based on the constraints section on the payload, the DMS found no matched machine.
When
- No technical dependencies.
Acceptance Criteria
Click to expand
- service-provider-dashboard readme file changed on gitlab and on gitbook.
Work Breakdown Structure (WBS)
| Task | Description | Duration | Status | Start Date | End Date | Comment |
|---|---|---|---|---|---|---|
| A | x Hrs | Done/In Progress |
Edited by Janaina Senna