search

juglab / n2v

This is the implementation of Noise2Void training.
Other
385 stars 107 forks source link

error in predicting model #135

Closed MHMoazezian closed 10 months ago

MHMoazezian commented 1 year ago

Hi I trained the model with a 2D array with 1 channel with shape (1600, 256,256,1) after generating patches. when I want to predict the trained model with another image with a shape of (256,256) I got this error : tile_overlap value for n_depth=6, kern_size=3, pool_size=2 not available. In your example just use an image as an input to the model.predict without any process!! Python 3.6.9 TensorFlow 2.4 n2v 0.3.1

veegalinova commented 1 year ago

Hi, can you please specify which of the examples you were following?

jdeschamps commented 10 months ago

I will close this issue due to lack of answers. Please do not hesitate to open a new one if you encounter problems.