This is part of a new effort to create a binary storage format for storing sparse matrices and other sparse data to disk.
Minutes from our meetings are available here (see also: previous minutes).
Here is a table listing the current tensor frameworks that support the format:
| Language | Framework | Status | Notes |
|---|---|---|---|
| C | binsparse-reference-c | under development | converts between binsparse and in-memory sparse matrices |
| C++ | binsparse-reference-impl | under development | converts between binsparse and in-memory sparse matrices |
| Julia | Finch.jl | complete | converts between binsparse and in-memory sparse matrices or tensors |
| Python | binsparse-python | under development | converts between binsparse and scipy.sparse matrices |
| Python | finch-tensor-python | complete | converts between binsparse and in-memory sparse matrices or tensors |
| C++ | TACO, binsparse-taco-parser | complete | converts between binsparse and in-memory sparse matrices or tensors |
The working version of the specification can be found under spec/latest/index.bs.
The spec is written in bikeshed – a variant of markdown. To render the spec locally:
- Install bikeshed (ideally in an isolated environment):
pipx install bikeshed - Call
bikeshed spec spec/latest/index.bs
To render the spec online, see api.csswg.org, or use the following commands:
cd binsparse-specification/spec/latest
curl https://api.csswg.org/bikeshed/ -F [email protected] > index.html
Rendered versions will generated for pull requests.