Skip to content

Latest commit

 

History

History
25 lines (15 loc) · 918 Bytes

README.md

File metadata and controls

25 lines (15 loc) · 918 Bytes

real-time-segmentation

1.This codes is related to real-time semantic segmentaion which run on the datasets of cityscapes,camvid and pascal voc2012.

Now,the codes can only work on the cityscapes,because I have not complete the codes about camvid and pascal voc2012 yet,but the work is still go on.

2.This repository contain FPN_ASPP,Nonlocal_Network,and BiseNetv1,and I will build BiseNetv2 in the future.

3.The command to train the code is:

python main.py --cuda --model "choicemodel" train --datadir data --num-epochs 30 --num-workers 4 --batch-size 4 --num-classes 20

dependency

pytorch>=0.4.0

approachs

1.FPN_ASPP

2.Nonlocal_Network

3.BiseNet(simple version)

result

image image