nicolay-r / AREkit

Document level Attitude and Relation Extraction toolkit (AREkit) for sampling and processing large text collections with ML and for ML
https://nicolay-r.github.io/arekit-page/
MIT License
58 stars 3 forks source link

What's new in 0.25.0 #536

Closed nicolay-r closed 7 months ago

nicolay-r commented 11 months ago

Support Batching for effecting imputing LLM into text processing pipelines

Previosly, the whole text processing pipeline was relying on the sentence / text part. Now we overcome that liimitation and therefore we can consider multiple sentences, formed in list i.e. batch. This step is so important for LLM, LM, neural networks, for which batching accelerates the performance. As the result, overall pipeline launching is expected to perform faster.

Sources collections are no longer going to be a part of AREkit :sparkles: Tha allow us to lightweight :feather: the overall framework and so that purely focus on data processing techniques

Flexibility and Performance Enhancements

nicolay-r commented 7 months ago

Published :partying_face: