Ktrain

New:
- support for building Question-Answering systems
- `textutils` now contains `paragraph_tokenize` function

Changed
- N/A

Fixed:
- resolved import issue with `textutils.sent_tokenize'

New:
- N/A

Changed
- `TransformerSummarizer` accepts BART `model_name` as parameter


Fixed:
- N/A

New:
- support for link prediction with graph neural networks
- text summarization with pretrained BART (included in 0.13.1 but not in 0.13.0)
- `bigru` method now selects pretrained word vectors based on detected language

Changed
- instead of throwing error, default to English if `detect_lang` could not detect language from batch of texts
- `layers` argument moved to `TransformerEmbedding` constructor
- enforce specific version of TensorFlow due to undocumented breaking changes in newer TF versions
- `AdamWeightDecay` optimizer is now used to support global weight decay. Used when user
excplictly sets a weight decay


Fixed:
- force re-instantiation of `TransformerEmbedding` object with `sequence_tagger` function is re-invoked

New:
- Added `max_momentum` and `min_momentum` parameters to `autofit` and `fit_onecycle` to control cyclical momentum

Changed
- Prevent loading errors of previously saved NERPreprocessor objects


Fixed:
- N/A

New:
- N/A

Changed
- Require at least TensorFlow 2.1.0 is installed in `setup.py` due to TF 2.0.0 bug with `lr_find`
- Added lower bounds to scikit-learn and networkx versions


Fixed:
- N/A

New:
- N/A

Changed
- N/A


Fixed:
- check and ensure AllenNLP is installed when Elmo embeddings are selected for NER