Ivis

Highlights:
* Improved training speed for numpy arrray inputs thanks to a faster triplet generator.
* Batched retrieval capabilities that makes ivis much faster when training on out-of-memory data that is retrieved in parallel.
* Improved performance when using Ivis with precompute=False option by using multi-threading when retrieving batches of KNN on-demand.
* Added deprecation notices for minor upcoming changes to API for consistency and adherence to sklearn API.

* Knn retrieval made more efficient by switching from multi-processing to multi-threading. Memory savings depend on OS and core count.
* Fixed issue where saved ivis models would attempt to load the index at the path they were saved with - this can't be relied on when the index is temporary and deleted after use.
* Fixed issue where Annoy Index metric parameter was not passed to an index that was loaded from disk.
* A few other things changed, including better error handling, cleaner code, and allowing for saving AnnoyKnnMatrix via pickle

* improved memory utilization during KNN retrieval
* AnnoyIndex is now removed from disk after running Ivis

Minor release addressing:

* Tensorflow 2.4 model save compatibility (82)
* Training/Inference batch size concordance

Major ivis release!