1.8.6 - Reverts torch dependency version to 2.0.1

What's Changed

• Revert torch dependency version to 2.0.1 by @antas-marcin in #70

Full Changelog: 1.8.5...1.8.6

1.8.5 - Fixed T5 model embedding generation

What's Changed

• Fix T5 model embedding generation by @antas-marcin in #67

Full Changelog: 1.8.4...1.8.5

1.8.4 - Added support for Bar-Ilan University NLP Lab models

What's Changed

• Add support for Bar-Ilan University NLP Lab models by @antas-marcin in #61
• Add build custom image step by @antas-marcin in #62

Full Changelog: 1.8.3...1.8.4

1.8.3 - Fixed downloading of nltk data into a local folder

What's Changed

• Fix downloading of nltk data into a local folder by @antas-marcin in #60

Full Changelog: 1.8.2...1.8.3

1.8.2 - Fixed asynchronous vectorization endpoint

What's Changed

• Added simple smoke tests to cover all of the endpoints by @antas-marcin in #57
• Vectorize asynchronously with ThreadPoolExecutor by @alexcannan in #56

New Contributors

• @alexcannan made their first contribution in #56

Full Changelog: 1.8.1...1.8.2

1.8.1 - Switch to Python 3.10 and torch 2.0.0

What's Changed

• Update supported models list by @antas-marcin in #53
• Switch to python 3.10 and update requirements to latest versions by @antas-marcin in #55

Full Changelog: 1.8.0...1.8.1

v1.8.0 - Fixed issue with T5 models on GPU

What's Changed

• Add support for google/flan-t5-base and google/flan-t5-large models by @antas-marcin in #51
• Sentence Transformers T5 GPU issue by @antas-marcin in #50
• Upgrade github action runner to ubuntu-latest-8-cores by @antas-marcin in #52

Full Changelog: 1.7.0...1.8.0

v1.7.0 - Added support for all-MiniLM-L6-v2 and multi-qa-distilbert-cos-v1 models

New Features

• Add support for all-MiniLM-L6-v2 and multi-qa-distilbert-cos-v1 models by @antas-marcin in #48

Full Changelog: 1.6.2...1.7.0

v1.6.2 - Fixed T2V_TRANSFORMERS_DIRECT_TOKENIZE setting

Fixes

• Fix T2V_TRANSFORMERS_DIRECT_TOKENIZE setting by @antas-marcin in #47

Full Changelog: 1.6.1...1.6.2

v1.6.1 - Added option to skip extra tokenization step during vectorization

New Features

• Optional direct text tokenization setting by @antas-marcin in #46

Full Changelog: 1.6.0...1.6.1