Deeppavlov

- A [relation extraction](http://docs.deeppavlov.ai/en/0.17.0/features/models/re.html) model
- A [ReCoRD model](https://github.com/deepmipt/DeepPavlov/blob/0.17.0/deeppavlov/configs/classifiers/superglue/superglue_record_roberta.json) is based on [Reading Comprehension with Commonsense Reasoning Dataset](https://sheng-z.github.io/ReCoRD-explorer/)

- A [distilled Russian BERT](https://github.com/deepmipt/DeepPavlov/pull/1465/files)
- New Pytorch-based [models](https://github.com/deepmipt/DeepPavlov/pull/1451) for classification, question answering and named entity recognition
- Removed some [configuration files and components](https://github.com/deepmipt/DeepPavlov/pull/1456)
- Small fixes and updates

- bert_as_summarizer, seq2seq_go_bot, hyperparameter optimization by neural evolution and all deeppavlov.deprecated components were removed
- [NER automodel](http://docs.deeppavlov.ai/en/0.15.0/apiref/models/torch_bert.html#deeppavlov.models.torch_bert.torch_transformers_sequence_tagger.TorchTransformersSequenceTagger)
- Сomponent-based config requirements generation
- Minor edits and fixes

- Fixed installation on python3.6

- [Intent Catcher](http://docs.deeppavlov.ai/en/0.14.0/features/models/intent_catcher.html) component
- [Support](http://docs.deeppavlov.ai/en/0.14.0/apiref/models/torch_bert.html?deeppavlov.models.torch_bert.torch_transformers_classifier.TorchTransformersClassifierModel#deeppavlov.models.torch_bert.torch_transformers_classifier.TorchTransformersClassifierModel) of the Hugginface Transformers for classification
- Go-Bot formfilling ([tutorial](https://github.com/deepmipt/DeepPavlov/blob/0.14.0/examples/gobot_formfilling_tutorial.ipynb))
- Entity Linking, Wiki Parser and KBQA as [separate components](http://docs.deeppavlov.ai/en/0.14.0/features/models/kbqa.html#advanced-using-entity-linking-and-wiki-parser-as-standalone-services-for-kbqa)
- Minor edits and fixes

- Hugging Face datasets [support](http://docs.deeppavlov.ai/en/docs-preview/apiref/dataset_readers.html#deeppavlov.dataset_readers.huggingface_dataset_reader.HuggingFaceDatasetReader)
- Go-Bot now requires only RASA-based configs to train its components (intent catcher, slot filler, dialogue state tracker also known as Go-Bot itself)
- Prometheus [metrics middleware](http://docs.deeppavlov.ai/en/0.13.0/integrations/rest_api.html#metrics) for REST web service
- KBQA models fixes
- Minor edits of the documentation