search

advaitsave / Multiclass-Semantic-Segmentation-CamVid

Tensorflow 2 implementation of complete pipeline for multiclass image semantic segmentation using UNet, SegNet and FCN32 architectures on Cambridge-driving Labeled Video Database (CamVid) dataset.
82 stars 41 forks source link

Confirm h5 weigths #9

Open dimitheodoro opened 4 years ago

dimitheodoro commented 4 years ago

Has anyone load the ''camvid_model_150_epochs.h5'' and try to predict a label?? i get terrible results... image my code: https://www.ppaste.org/JduAJyKTY#