Config class

implementation of bisenet v2

Read training examples from tfrecords for nsfw model

init lanenet multi gpu trainner

VGG 16 based fcn net for semantic segmentation

implementation of context embedding module in bisenetv2

implementation of gather and expansion module in bisenetv2

implementation of guided aggregation module in bisenetv2

Instance segmentation result cluster

implementation of segmentation head in bisenet v2

implementation of stem block module

Base model for other specific cnn ctpn_models

LaneNet backend branch which is mainly used for binary and instance segmentation loss calculation

Convert raw image file into tfrecords

LaneNet frontend which is used to extract image features for following process

lanenet post process for lane generation

init lanenet single gpu trainner