v0.0.7

1. Bugfix: update dependency version of longling to solve the problem using torch light_module with cuda

v0.0.6

1. Bugfix: update dependency version of longling

PyBaize v0.0.5

1. mxnet: add split_and_load
2. add bucketing

PyBaize v0.0.4

1. utils: add io relevant including json_load, as_out_io, as_io
2. utils: add path relevant including path_append
3. torch: add rnn layer: get_state_shape, get_init_state
4. torch: add erase-add gate
5. torch: add functional.utils including tensor2list, batch_select
6. torch: add functional.seq including pick, length2mask, get_sequence_mask, mask_sequence
7. torch: add functional.rnn including get_rnn_state_shape, get_rnn_init_state
8. add analysis relevant: get_best
9. add get_epoch_params_filepath
10. add Timer, as_list, config_logging, iterwrap
11. update dependencies: longling
12. add constants: CFG_JSON, RESULT_JSON
13. add new function: get_params_filepath

PyBaize v0.0.3

1. add configuration parser
2. add save_params and load_net for torch and mxnet

PyBaize v0.0.2

1. Add light_module for both torch and mxnet
2. add modelzoo for mxnet, including attention, highway, pairwise loss and TextCNN and TransE

PyBaize v0.0.1.post2

Publish

PyBaize v0.0.1

Initialization