* make the pix2pix support multi gpu trainning

* Update Image2Image.py

* Update Image2Image.py

* Update Image2Image.py

* make dataflow idiomatic python container objects (fix #869)

* use ABCMeta and fix tqdm

* python2 compatible version

* fix

* some more replacement

* rollback changes of _reset_df_and_get_size

* fix unrelated pbar issue

* revert the __len__ changes on InputSource

* use svg

I can't look at it anymore.

* Update README.md

use gitter brand color #ed1965 or #46bc99

* add flownet2-S inference example

* fix readme

* flake8 fix

* fix flownet-s, flownet-c scaling

* uuu

* Small simplifications

* Add centercrop in inference. Update readme