Benchmark files for NLP problem

Benchmark files for NLP problem for testing WBF 1D case.

Benchmark files

Benchmark

Speed up + New Features

• Added many input data checks in all methods: set to 0.0 coords < 0, set to 1.0 coords > 1, swap cases where x2 < x1 and y2 < y1, remove zero area boxes.
• Added numba @jit(nopython=True) for critical functions (NMS, NMW and WBF). Speed up around x2 times (tested on example_oid.py).
• Added support for 3D boxes with function weighted_boxes_fusion_3d

Bug fix

Fixed bug with incorrect work of WBF and NMW algorithms if provided boxes was unsorted and skip_box_thr was larger than 0.0.

First stable version

v1.0.0

Merge remote-tracking branch 'origin/master'

Initial release + test data

v1.0

Init