Update map() and optimize() documentation #264
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #264 +/- ##
=====================================
Coverage ? 78%
=====================================
Files ? 34
Lines ? 4874
Branches ? 0
=====================================
Hits ? 3785
Misses ? 1089
Partials ? 0 |
senarvi
changed the title
senarvi
force-pushed
the
update-map-and-optimize-documentation
branch
5 times, most recently
from
a3b5f6d
to
4446d77
Compare
senarvi
force-pushed
the
update-map-and-optimize-documentation
branch
from
4446d77
to
16350bf
Compare
for more information, see https://pre-commit.ci
|
Hi @senarvi, thanks for opting to contribute to litdata. Btw, Can you please convert your PR into draft until it's ready for review! |
|
@deependujha how do I convert a PR into a draft? |
|
Does anyone have an idea why the Windows and MacOS tests fail with "Numpy is not available"? |
Just click on convert to Draft. I'll try attempting the failing tests soon. |
|
Thanks @deependujha , I haven't noticed that link! That will be useful in the future. From my behalf this PR is ready. I think anything in this PR is not responsible for breaking the tests, since this doesn't include functional changes. So better to leave this open, or what do you think? |
|
sure. I'm looking into it. Thanks for contributing. |
There was a problem hiding this comment.
All failing tests were bcoz of numpy. But, recently support for numpy>=2 was added.
cc @tchaton
Fixes documentation related to #260 .
Adds notes that the types and list sizes must not changed between samples, when calling
optimize().When looking at the code, I noticed that the documentation of the inputs seems to be outdated. AFAICS, inputs can be any sequence or a streaming dataloder, so I updated the type annotation and removed this from the docstring: "Each input should contain at least a valid filepath". Let me know if I missed something, or feel free to make changes in the PR.
My code editor also removed unnecessary whitespace. I guess that's good.