-
Notifications
You must be signed in to change notification settings - Fork 5
Project subtensor milestones
While I'll be following the rough outline of two phases, these milestones are a little fine-grained to be at most 2-3 weeks as I tend to work better with shorter deadlines.
Major Milestone | Milestone | Issues | Start Date(inclusive) | Due date(inclusive) |
---|---|---|---|---|
Phase-1 | Initial subtensor implementation |
link | June 7, 2021 | July 18, 2021 |
Phase-1 | First evaluation | - | July 12, 2021 | July 18, 2021 |
Phase-2 | Auxiliary-types implementation | link | July 19, 2021 | August 15, 2021 |
The milestone defines the work done to implement the support of subtensor
's in Boost.uBLAS the primary aim of my project. More information about the work done can be found in the here.
Definition of Done: Completing this milestone would signify that the implementation of subtensor
has been created that acts as a view of a tensor
and should provide tensor functions and be treated as a tensor.
Auxiliary types such as span
and slice
are especially the primary way a user would create and interact with subtensor
s. Such types increase the expressiveness of the library. They were implemented during GSoC'19. During this phase, I would improve such types and integrate them with my implementation of the subtensor
.
Definition of Done: Auxiliary types for span and slicing, strides, etc. of tensor
and subtensor
This milestone is basically to shadow the actual project milestone mandated by google.
The implementation needs to be sufficiently tested and refactored, and improved, and incorporate any changes. Documenting and preparing the work to be finally merged into Boost.uBLAS.
Definition of Done: Final subtensor
implementation and work done during GSoC and write a final project report.
We both would like to thank our mentor Cem Bassoy for his constant support and help in achieving our milestones. From conducting regular meetings and providing indispensable information and suggestions, to the helpfulness in always answering our queries we always find him extremely helpful and his contribution towards our projects is massive. We are grateful for the time and efforts he devoted towards our projects. We would also like to thank Google for the Google Summer of Code Programme, without which all these wouldn't be possible. We would like to take the opportunity to thank the members of the of Boost Community for letting us be a part of this wonderful journey.