Arenops

Trained on cityscapes with 4 GPUs and batches==20 and lr==0.0006 in config.py in deeplabv3+voc, get **mIOU 24.867% after 32 epochs**. Loss is remain 0.3-0.6 and no longer reduce. @YudeWang any...

Nice work! Will you release your training scripts?

If we define x as a big 2-D dimensional matrix ( like a image with resolution 1920*1080 ), the GPU memory will increase quickly when optimizing.