Viktor Gal
Viktor Gal
as a quick has i've tried moving every IO logic into a separate worker thread that works on a queue to process all the IO operations and then fill out...
apparently indirection of indirection should solve this problem: http://blog.hostilefork.com/qt-model-view-different-threads/ meaning everything needs to be signaled... although since now i've moved everything out of `BrowseFeature` into `FolderItemModel`, i guess that indirection...
so, i'm hoping that this is the Qt standard solution for the problem i was trying to fix. unfortunately this solution results in an as bad as experience that this...
It’s all good. Imo this PR is ready as is and should be good to merge. Unless you want something specific to be amended. Note with the last feedback on...
@poelzi rebased and fixed. next step? as this ought to happen in the future so would be good to have a direction of this PR.
i'll try to refactor & cleanup based on @ywwg comments this week
@iglesias @manish7294 has some question about LMNN implementation in shogun :)
note i have tried the same code with `datasets` version 2.4.0, the outcome is the very same as described above.
@lhoestq thnx for getting back to me! i've tested the suggested method, but unfortunately the memory consumption is the very same: ``` from datasets import Dataset, Features, Array2D, Array3D, load_from_disk...