search

MeiShaohui / 3DCAE-hyperspectral-classification

Unsupervised Spatial-Spectral Feature Learning by 3-Dimensional Convolutional Autoencoder for Hyperspectral Classification
85 stars 31 forks source link

Different problem in train, mode = test #2

Open davodogster opened 5 years ago

davodogster commented 5 years ago

When i run train.py, mode = test, I get this error:

ValueError: Layer #2 (named "PReLU1"), weight <tf.Variable 'PReLU1/alpha:0' shape=(201, 143, 143, 24) dtype=float32_ref> has shape (201, 143, 143, 24), but the saved weight has shape (201, 3, 3, 24).

mcvarer commented 4 years ago

yes i have a same problem

komorebi404 commented 4 years ago

I have the same problem. Have you solved it?

ZZYhengiyou commented 2 years ago

have you solved this problem? I have the same problem.

Ritatanz commented 2 years ago

Change the 'feature_model = DCAE_fea()' to ' feature_model = DCAE()' in the class ModeTest. We just need to make sure we use the same CAE model for the training and testing parts.

ZZYhengiyou commented 2 years ago

thanks for your reply! I will give it a try