search

feevos / resuneta

mxnet source code for the resuneta semantic segmentation models
Other
127 stars 30 forks source link

How many images did you use to train? #4

Closed thimabru1010 closed 4 years ago

thimabru1010 commented 4 years ago

I am trying to implement the multitasking version in keras on tf 2.0 and wanted some parameters to compare.

feevos commented 4 years ago

Hi, I used all the training data provided by the ISPRS, for training and validation. The validation dataset was ~10% of the total area. Importantly (perhaps), I extracted validation data from each tile, i.e. I did not keep some tiles exclusively for validation.