Upload auxilary files for decoding

* add AP figures

* Small fix for doc.

* Refine evaluation code.
* Clean code.

* add face_detection/README_cn.md for pyramidbox

* Guide for face detection.
* Follow comments.

Add README for Transformer

Add Chinese README for DeepASR

Adapt decoder to the new net config

Guide for non-image fluid models

fix format of deepqnetwork doc

Fix the max length in Transformer to count start and end tokens in

* Introduction for models about computer vision.
* Follow comments.

Add Chinese version of README and provide saved models

* Speed data processing by multi-threads/multi-process.
* Add profiling scripts.
* Use depthwise transposed conv2d.

* add more pyramid

Add transformer beamsearch decoder

Apply exponential decay on learning rate