Releases: activeloopai/deeplake
Releases · activeloopai/deeplake
v4.0.3
v3.9.29 🌈
v4.0.2
v3.9.28 🌈
v4.0.1 🌈
🧭 What's Changed
- Up to 5x improvement on dataset open time for large datasets
- Ability to interrupt long running query with Ctrl-C
- Added support for tag checkout in TQL FROM
- Reduced binary size by ~2Mb
- Fixed clustered index search on small data
- Fixed deeplake.convert bug with text columns
v3.9.27 🌈
v4.0.0 🌈
🧭 What's Changed
Major release of Deep Lake
- Eventual Consistency: Enabling concurrent workloads. Read more here.
- Faster Installation: 5x faster setup by removing all dependencies except NumPy.
- Enhanced Performance: Up to 10x faster reads/writes due to migrating low-level code to C++.
- Cross cloud queries with JOIN operations and User-Defined Functions
Simplified API with unified documentation, better data typing, and async support. - Switched to Apache 2.0 license
As a major release, the 4.0 API is not compatible with 3.x. Full documentation is available at docs.deeplake.ai including a v3 to v4 conversion guide.
Note: Open source code for 4.0 will be fully moved into this repo in the coming weeks
Bugfixes on 3.x should be made against the v3 branch
v3.9.26 🌈
v3.9.25 🌈
v3.9.24 🌈
🧭 What's Changed
- bump libdeeplake version to 0.0.146 (#2962) @activesoull
- strict numpy requirement (#2961) @activesoull
- bump libdeeplake version to 0.0.146 (#2960) @activesoull
- added pyproject for poetry support (#2959) @activesoull
- strip dataset name (#2956) @azat-manukyan
- Retry connection errors on credentials requests (#2957) @nvoxland-al
- local storage (#2954) @activesoull
- Process hub path (#2953) @activesoull
- Upgrade pillow (#2952) @activesoull
- Fix partial read (#2951) @activesoull
- Invalid corrupt sample message (#2950) @activesoull