v0.0.10
Pre-release
Pre-release
Dear friends, PyPOTS v0.0.10 is here! 🥳
In this version, we
- added support for training models on large datasets with a lazy-loading strategy. I believe this feature is necessary for our users who want to apply PyPOTS in some industrial scenarios where massive datasets are common;
- updated the input format. To be compatible with lazy-loading large data from file, the input of function
fit()
now has been changed fromX
totrain_set
.train_set
has to be a dict like {"X": X, "y": y} or a path string locating a HDF5 file. - replaced
print
withlogging
, which can show necessary printing information more structured; - fixed some inconspicuous and non-fatal bugs;
- added test cases to top up the code coverage;
- synchronously release PyPOTS in the channel
conda-forge
on Anaconda, now you can install with conda;
Still, this is a pre-release before the first official version 0.1, and this probably is the last pre-release before v0.1
. I'm still working on the docs, tutorials, and new models. Ping me if you have any suggestions or comments! Many thanks, my friends!
What's Changed
- merge branch dev into main by @WenjieDu in #29
- Update README.md and requirement.txt by @WenjieDu in #33
- Merge
dev
intomain
by @WenjieDu in #37 - Merge
dev
intomain
by @WenjieDu in #38 - Merge
dev
intomain
by @WenjieDu in #39
Full Changelog: v0.0.9...v0.0.10