Texts were mentioned as-it-is since for now we are lack of training data to reproduce the same model for English language. In terms of understanding list: 1

Texts were mentioned as-it-is since for now we are lack of training data to reproduce the same model for English language. In terms of understanding list: 1. Yes it is. 2. Entity tagging model, based on DeepPavlov, is a part of ARElight, not AREkit. 3. AREkit implements and provides tools for BERT data preparation only; BERT-related service is a part of the ARElight. Overall idea and understanding is correct. In terms of the AREkit, the latter proposed as a data-preparation toolkit. However, at present, it additionally provides a Tensorflow-based neural networks zoo in, adopted for sentiment relation extraction task https://github.com/nicolay-r/AREkit/tree/0.22.0-rc/arekit/contrib/networks/context/architectures https://github.com/nicolay-r/AREkit/tree/0.22.0-rc/arekit/co...