GeorgeSeif / Semantic-Segmentation-Suite

Semantic Segmentation Suite in TensorFlow. Implement, train, and test new Semantic Segmentation models easily!
2.51k stars 878 forks source link

Not working for input crop size other than 512 X 512 #195

Open suddu22 opened 5 years ago

suddu22 commented 5 years ago

Hi I am trying to run any model on the CamVid Dataset with 13 classes with crop size of 360 X 480 (anything other than the default). It gives incompatible size error for any model. Is there something I am doing wrong or is the code compatible only with size 512 X 512?

ryohachiuma commented 5 years ago

Which kind of model do you want to run? Input image size can be divided by 32 for some models. You have to read the paper and figure out the restriction of input image size.

K-M-Ibrahim-Khalilullah commented 4 years ago

@kemangjaka Same problem for me on the FC-DenseNet103. Please check.. https://github.com/GeorgeSeif/Semantic-Segmentation-Suite/issues/230

ryohachiuma commented 4 years ago

That error you have is about GPU memory. You don't have enough GPU memory to feed-forward that size of the image. I think you have to change the larger GPU or make the image smaller.

yucheng-china commented 4 years ago

Which kind of model do you want to run? Input image size can be divided by 32 for some models. You have to read the paper and figure out the restriction of input image size.

Hi, brother, where is the paper, i want to read them