Ktrain

New:
- N/A

Changed:
- N/A


Fixed:
- store a local version of `stellargraph` to prevent it from installing `tensorflow-cpu`
and overriding existing `tensorflow-gpu` installation

New:
- Support for node classification in graphs with `ktrain.graph` module

Changed:
- N/A


Fixed:
- N/A

New:
- N/A

Changed:
- N/A


Fixed:
- Call `reset` before `predict_generator` for consistent ordering of `view_top_losses` results
- Fixed incorrect reference to `train_df` instead of `val_df` in `texts_from_df`

New:
- All `fit` methods in *ktrain* now accept `class_weight` parameter to handle imbalanced datasets.

Changed:
- N/A


Fixed:
- Resolved problem with `text_classifier` incorrectly using `uncased_L-12_H-768_A-12` to build BERT model
instead of `multi_cased_L-12_H-768_A-12` when non-English language was detected.
- Fixed error messages releated to preproc requirement in `text_classifier`
- Fixed test script for multingual text classification
- Fixed rendering of Chinese in `view_top_losses`

New:
- N/A

Changed:
- N/A


Fixed:
- Fix problem with `text_classifier` incorrectly using `uncased_L-12_H-768_A-12` to build BERT model
instead of `multi_cased_L-12_H-768_A-12` when non-English language was detected.

New:
- Added multilingual support for text classification.
- Added experimental support for tf.keras. By default, *ktrain* will use standalone Keras.
If `os.environ['TF_KERAS']` is set, *ktrian* will attempt to use tf.keras.
Some capabilities (e.g., `predictor.explain` for images) are not yet supported for tf.keras

Changed:
- When BERT is selected, check to make sure dataset is correctly preprocessed for BERT


Fixed:
- Fixed `utils.bert_data_type` and ensures it does more checks to validate BERT-style data