Thanks.
We will figure out how to push the code on MLCask onto Apache SINGA.
MLCask is built on top of Apache SINGA and ForkBase for handling
component evolution
and model-data provenance. We will have to figure out how to remove the
dependency
on ForkBase, as it is a closed source at this moment -- one possible
way, is to use
Rocksdb and build the necessary features (found in ForkCloud).
https://arxiv.org/pdf/2010.10246.pdf
regards
beng chin
On 2021-03-07 11:00, Wang Wei wrote:
Dear all,
Here is a draft of the report. Please feel free to add your comments by
10
March.
I would also like to get your ideas about the next release.
I think we should continue enhancing the distributed training and then
add
new features, e.g., serverless inference.
Thank you!
Regards,
Wei
-----------------------------
## Description:
The mission of Apache SINGA is the creation and maintenance of software
related
to a distributed deep learning platform
## Issues:
N.A
## Membership Data:
Apache SINGA was founded 2019-10-16 (a year ago)
There are currently 22 committers and 16 PMC members in this project.
The Committer-to-PMC ratio is roughly 3:2.
Community changes, past quarter:
- No new PMC members. Last addition was Chris Yeung on 2020-04-17.
- No new committers. Last addition was Zhang Zhaoqi on 2020-07-01.
## Project Activity:
Recent releases:
3.1.0 was released on 2020-10-30.
3.0.0 was released on 2020-04-18.
The major activities were fixing some bugs and issues of V3.1.0.
The major features of the next release are to be discussed.
## Community Health:
Multiple committers were moving to new roles (companies). The
development
progress was slowed down compared to the last report. The community
needs
to attract new committers and may restructure the core development
team.
Commit activity:
27 commits in the past quarter (-76% decrease)
3 code contributors in the past quarter (-72% decrease)
GitHub issues:
4 issues opened on GitHub, past quarter (-60% decrease)
6 issues closed on GitHub, past quarter (-62% decrease)
GitHub PR activity:
30 PRs opened on GitHub, past quarter (-53% decrease)
34 PRs closed on GitHub, past quarter (-58% decrease)
