0.6.4

We now support faster inference on M1 macs natively out of the box!

0.6.2

Add direct support for matrouska models and make sentence-transformers a core dependency.

0.6.1

Fixed a keras import bug.

0.6.0

This version introduces breaking changes by splitting the finetuning submodule into tuners and learners. These are now the available components.

# Finetuning components 
from embetter.finetune import FeedForwardTuner, ContrastiveTuner, ContrastiveLearner, SbertLearner

The docs also got a nice update as a result. We hope folks enjoy and enjoy experimenting!

0.5.3

Add support for lite text embedding models.

0.5.2

Added support for a difference model.

0.5.1

New documentation pages and support for Gensim/KerasNLP.

0.5.0

Introduced a breaking change: external embedding providers will now require an environment key. Feels safer to force users to use that.

0.4.1

Added a whole bunch of utilities #71.

0.4.0

Add support for caching the embeddings.