Hsuan
Hsuan
Hey @ogail, Since by default is to load pre-trained model and keep finetuning on it. However, the pre-trained cityscapes has 19 classes, while your dataset has only 1. You can...
Before that, I want to know what your dataset look like, can you show some examples? If there is only one class, it doesn't need to train anymore, am I...
It make sense to me. For this case, I think it's difficult to learn to detect obstacles, since the obstacles contain several different objects. Hence, I think you can restore...
Hey @aliericcantona, it seems that the checkpoint is missing, so you cannot load the pre-trained weights. You can refer to #45.
@aliericcantona, I have updated the model but the output results not, so the results would be a little different.
Hey guys, I would suggest that you can use`tf.Dataset` API to validate the model during training. I will try to update and clean the code in recent days. Thank @BCJuan...
Hey @limitime , Since my indoor model is used to robot navigation, it can only recognize the walls, floors, and ceilings. If you need, I can provide you.
@limitime, okay, I'll upload the checkpoint and weight soon. Could you tell me what you are working with XD? I'm curious about where you can use indoor segmentation.
@limitime Here is the model weight: https://drive.google.com/drive/folders/1zHFCXRyaW7OFn9zA41_c7Hfnl1HMhgK_?usp=sharing Remember to change the class number from 19 to 27.
@limitime, sorry for late reply. You don't need change the IMG_MEAN, and I don't change it when training, either.